Founder-CTO & AI Product Engineer
+Seoul, South Korea | untilhamza@gmail.com | (+82) 10-9899-9793 | linkedin.com/in/untilhamza | github.com/untilhamza
+Founder-CTO and AI product engineer who builds products where domain workflows, AI systems, and monetization meet. Currently focused on HoverNotes, a revenue-generating AI video learning product with over $10K/month revenue and around 20K Chrome/Edge installs, and Quest, a safari-operator SaaS platform I co-founded as CTO. Strongest at turning awkward manual workflows into trusted software: understand the domain, design the product and business loop, build the full stack, instrument behavior, and iterate in production. Current xAI Human Data contract work adds hands-on signal in software-engineering model evaluation. Fluent in Korean and English.
+Product & Business Systems: 0-to-1 SaaS, pricing strategy, subscriptions, trials, paywalls, funnel analysis, SEO/i18n, customer-facing reliability
+AI/LLM Systems: multimodal AI workflows, AI note generation, coding-model evaluation, RAG, LangChain, LangGraph, Whisper, OpenAI API
+Full-Stack Platform: Next.js, React, TypeScript, Python/FastAPI, Node.js/Express, PostgreSQL, Supabase, Drizzle ORM, MongoDB, Redis
+Workflow & Document Systems: browser extensions, Obsidian/local-first workflows, rich editors, PDF generation, email delivery, Cloudflare R2
+Infrastructure & Ops: Vercel, AWS Lambda/S3/EventBridge, Docker, Hetzner, CI/CD, PostHog, Sentry, Stripe, Razorpay
+JonjaBird | Founder / product operator | 2026 - Present: Designed a physical-digital mail and support service with pricing, client ledgers, receipt/deposit confirmation PDFs, account statements, and a future sponsor/admin platform spec.
+Ugandan Community in South Korea | Vice President | 2026 - Present: Support governance, embassy-facing communication, partnership strategy, and community engagement materials for the Ugandan diaspora association in South Korea.
+Korea University, Seoul: B.S. Electrical and Electronics Engineering, Mar 2019 - Feb 2023, GPA 4.02/4.5. Magna Cum Laude; Global Korea Scholarship and Ugandan Government Scholarship recipient.
+Co-authored "Responsible Music Genre Classification using Interpretable Model-agnostic Visual Explainers," Springer Nature Computer Science.
+Contributed to developer tools and AI/open-source ecosystems including LangChain, Vercel AI SDK, Editor.js, Redux Toolkit, MemGPT, Pytube, FastAPI, and related projects. Published widely used technical answers and references for WebRTC, speech-to-text, AI application patterns, and production integration issues, bringing practical debugging experience back into product work.
+Technical Founder, CTO & Senior AI Product Engineer
+Seoul, South Korea | untilhamza@gmail.com | (+82) 10-9899-9793 | linkedin.com/in/untilhamza | github.com/untilhamza
+Technical founder, CTO, and senior AI product engineer focused on HoverNotes and Quest, with 4+ years building AI-powered education, productivity, travel, and workflow systems. Founded HoverNotes, now over $10K/month revenue with around 20K Chrome/Edge installs, and co-founded Quest as CTO, leading a 3-person team building safari quotation software and PDF infrastructure. Also contracted with xAI Human Data on software-engineering model evaluation. Strong in TypeScript/React/Next.js, Python/FastAPI, browser extensions, LLM systems, payments, analytics, document generation, and production reliability. Fluent in Korean and English.
+AI/LLM: OpenAI API, Whisper, LangChain, LangGraph, RAG, vector search, prompt engineering, multimodal AI workflows
+Frontend/Product: Next.js, React, TypeScript, React Native, Tailwind CSS, Lexical, Chrome/Edge extensions, Obsidian/local-first workflows
+Backend/Data: Python, FastAPI, Node.js, Express, PostgreSQL, Supabase, Drizzle ORM, MongoDB, Redis
+Infrastructure/Ops: AWS Lambda, S3, EventBridge, Docker, Vercel, Cloudflare R2, Hetzner, CI/CD, PostHog, Sentry
+Business Systems: Stripe, Razorpay, webhooks, subscriptions, trials, pricing, credits, PDF/document generation
+Languages: Korean (TOPIK 5), English, Luganda
+Korea University, Seoul: B.S. Electrical and Electronics Engineering, Mar 2019 - Feb 2023, GPA 4.02/4.5. Magna Cum Laude; Global Korea Scholarship and Ugandan Government Scholarship recipient.
+Co-authored "Responsible Music Genre Classification using Interpretable Model-agnostic Visual Explainers," Springer Nature Computer Science.
+Technical Founder @ HoverNotes
+Founder-CTO & AI Product Engineer
Seoul, South Korea
-news: false # includes a list of news items -latest_posts: false # includes a list of the newest posts -selected_papers: false # includes a list of papers marked as "selected={true}" -social: true # includes social icons at the bottom of the page +news: false +latest_posts: false +selected_papers: false +social: true --- - +Right now I spend most of my time on two products: - +- **[HoverNotes](https://www.hovernotes.io)**, a privacy-first AI video note-taking product that turns online lectures, tutorials, and course videos into structured notes users can keep in their own Obsidian/Markdown workflows. I founded and built the product across Chrome/Edge extension workflows, a Next.js web app, realtime video-in notes-out AI workflows, payments, analytics, localization, SEO, and production operations. As of June 2026, HoverNotes is doing over $10K/month in revenue with around 20K installs across Chrome and Edge. +- **[Quest](https://questheaven.com)**, a safari quotation platform for tour operators. I co-founded Quest as CTO and lead a 3-person engineering/product team building operator dashboards, admin tools, travel requests, itinerary planning, pricing, branded PDF generation, quote email delivery, content libraries, billing, and release workflows. - +I also contract with **xAI Human Data** on software-engineering model evaluation, reviewing AI-generated code across languages and domains to help improve coding-model quality. - +Before this, I was a **Lead AI Engineer at Bebridge/Slid**, a 50K+ MAU EdTech platform, where I took on principal-level AI engineering work: setting technical direction, solving hard AI/product problems, and shipping LLM-powered auto-notes, AI chat, semantic search, transcript correction, mobile AI workflows, real-time transcription, cross-platform auth/payments, and multi-format exports. - +#### Engineering Focus - +- **AI product systems:** LLM workflows, multimodal AI, RAG, coding-model evaluation, AI note generation, transcript correction, and learning tools. +- **Full-stack platforms:** TypeScript, React, Next.js, Python/FastAPI, Node.js/Express, PostgreSQL, Supabase, Drizzle ORM, MongoDB, and Redis. +- **Workflow infrastructure:** browser extensions, local-first Obsidian workflows, rich editors, PDF/document generation, email delivery, Cloudflare R2, Stripe, Razorpay, Vercel, AWS, Docker, PostHog, and Sentry. -#### Welcome to My Personal Website +My engineering philosophy is simple: put users at the center, move fast but measure real behavior, build the whole loop from product to reliability, and stay calm under ambiguity. -I am a Technical Founder and AI Engineer specializing in full-stack + AI development with 4+ years of hands-on experience building revolutionary educational technology. Currently founding HoverNotes. +#### Background -#### Education & Professional Career +I graduated Magna Cum Laude from [Korea University](https://www.korea.edu/mbshome/mbs/en/index.do) with a B.S. in Electrical and Electronics Engineering and a 4.02/4.5 GPA. I received both the Ugandan Government Scholarship and the [Global Korea Scholarship](http://www.niied.go.kr/user/nd35203.do). -I am a proud recipient of the Ugandan Government Scholarship (2017) and the [Global Korea Scholarship](http://www.niied.go.kr/user/nd35203.do) (2018 - 2023). I graduated cum laude in Electrical Engineering from [Korea University](https://www.korea.edu/mbshome/mbs/en/index.do), Seoul South Korea (class of 2023) with a 4.02/4.5 GPA. +I co-authored a Springer Nature Computer Science paper on responsible AI and interpretable machine learning, and I have contributed to developer tools and AI/open-source ecosystems including LangChain, Vercel AI SDK, Editor.js, Redux Toolkit, MemGPT, Pytube, FastAPI, and related projects. -**Current Roles:** -- **Technical Founder @ HoverNotes**: Building privacy-first AI video learning platform with universal video integration, local Obsidian storage, and 45+ language support +I also serve as **Vice President of the Ugandan Community in South Korea**, supporting governance, embassy-facing communication, partnership strategy, and community engagement. -**Previous Experience:** -- **Fullstack Engineer @ GoGymi (Jan-Sep 2025)**: Architected and built TexTutor AI-powered educational platform featuring multi-language grammar correction system, Stripe billing integration with subscription management, real-time collaborative editing with LiveBlocks, comprehensive authentication system, email infrastructure with multi-language support, and responsive React/Next.js dashboards for teachers and students (pre-launch) -- **Principal AI Engineer @ Slid (2022-2025)**: Pioneered revolutionary features including Auto Notes System (60% cost optimization), Smart Live Text (25% premium subscription increase), AI Sliddy chatbot, and multi-agent transcript correction systems +#### Connect -**Research:** Co-authored paper in SN Computer Science (Springer Nature) on responsible AI and explainable machine learning. +I’m open to senior/principal engineering conversations, technical advisory, AI product collaborations, privacy-first AI, coding agents/model evaluation, and workflow automation. - - -#### None profits - -I founded Jenga Code Labs in my final year at Korea University. Our mission is _Empowering foreign students in Korea to overcome social and professional barriers by equipping them coding skill to enhance their career trajectories_. - - - - - - - -#### Let's Connect - -Interested in LLMs, software innovation, or just a friendly chat about the latest in tech? Reach out to me at [untilhamza@gmail.com](mailto:untilhamza@gmail.com) or connect with me on [LinkedIn](https://www.linkedin.com/in/untilhamza) or [GitHub](https://github.com/untilhamza). - -#### Open Source - -- 🤝 I've contributed to the following open source projects :- - - [Langchain](https://github.com/langchain-ai/langchain) A framework for developing applications powered by language models.[[discord]](https://discord.com/invite/5Fgux4em9W) - - [MemGPT](https://github.com/cpacker/MemGPT) Long term memory for LLM agents [[paper]](https://arxiv.org/abs/2310.08560) [[discord]](https://discord.com/invite/9GEQrxmVyE) - - [Vercel AI](https://github.com/vercel/ai) Library for building AI-powered user interfaces. - - [Redux Toolkit](https://github.com/reduxjs/redux-toolkit) The standard way to write Redux logic[[discord]](https://discord.com/invite/0ZcbPKXt5bZ6au5t) - - [EditorJs](https://github.com/editor-js) A block-style editor with a universal JSON output - - [MDN](https://github.com/editor-js) Official documentation of the web - - [Pytube](https://github.com/editor-js) Lightweight library for downloading YouTube Videos - -#### TL;DR - -- 🌍 I'm based in Seoul, South Korea - - -- 📫 You can reach me at [untilhamza@gmail.com](mailto:untilhamza@gmail.com) -- 🚀 I'm currently working on a [privacy first video learning technology ](http://www.hovernotes.io) -- :microscope: My research interests are in **Generative AI**, **Large Language Models** and **Responsible AI**. -- :briefcase: [My Resume](/assets/pdf/2025-10-24-Hamza-Kyamanywa-Resume.pdf) - - -- ⚡ Fun fact **I speak a language from every continent!** 🌎 - - - -### Skills - -- **Languages:** TypeScript, Python, JavaScript, C++ -- **Frontend:** React 19, Next.js 15, React Native, Chrome Extensions (Manifest V3), Tailwind CSS, Bootstrap -- **Backend:** FastAPI, Node.js, NestJS, Express, Django, Flask -- **AI/ML:** LangChain, LangGraph, GPT-4, Claude, Gemini, PyTorch, Keras, SHAP, LIME, Vector Databases -- **Real-time:** WebRTC, WebSocket, Speech-to-Text (STT), Real-time Transcription -- **Databases:** PostgreSQL, MongoDB, Redis, MySQL, Firebase, Supabase, Vector DBs -- **Infrastructure:** Docker, Kubernetes, AWS, GCP, Multi-region Deployment, CDN -- **Browser APIs:** File System Access API, Clipboard API, Screen Capture API, Media APIs -- **Architecture:** Microservices, Multi-agent Systems, Event-driven, Serverless -- **Other:** Git, Agile/SCRUM, Anti-detection Systems, Performance Optimization - - - -### My GitHub Stats - -Quest is a safari quotation platform for tour operators. I co-founded Quest as CTO and lead a 3-person engineering/product team building the platform: operator dashboards, admin tools, travel requests, itinerary planning, pricing, branded PDF generation, quote email delivery, content libraries, billing, and release workflows.
+ +Tour operators often build quotes through spreadsheets, scattered assets, manual document editing, and repeated back-and-forth with travelers. Quest turns that workflow into structured software: requests become itineraries, pricing becomes traceable, and final proposals become branded PDF and email outputs.
+ +Next.js 16, React 19, TypeScript, Tailwind CSS, PostgreSQL/Supabase, Drizzle ORM, NextAuth v5, Resend, Stripe, PostHog, Cloudflare R2, TypeScript/Express, Handlebars, Puppeteer, Docker, Hetzner, and Vercel.
+I contract with xAI's Human Data software-engineering track to create and review high-quality coding data for training, benchmarking, and improving large language models.
+ +This work sits close to the quality layer of AI coding systems: identifying which outputs actually satisfy real software-engineering tasks, why they work or fail, and how evaluation evidence should be grounded in code behavior rather than vibes.
+Infrastructure Design, Video Processing, Docker, Kubernetes, Performance Optimization, Anti-Detection Systems, Global Deployment, System Monitoring
- \ No newline at end of file + diff --git a/_projects/1_project.md b/_projects/1_project.md index 3bf731d..cd87810 100644 --- a/_projects/1_project.md +++ b/_projects/1_project.md @@ -1,51 +1,32 @@ --- layout: page title: HoverNotes -description: Privacy-first AI video learning platform eliminating context-switching -img: # assets/img/12.jpg +description: Privacy-first AI video note-taking product with 20K Chrome/Edge installs and over $10K/month revenue +img: importance: 1 category: work ---Founded and leading technical development of HoverNotes, a groundbreaking privacy-first AI video note-taking platform that eliminates context-switching in video-based learning. The platform features a sophisticated three-tier architecture (browser extension + web app + AI backend) enabling note-taking on any video platform while maintaining complete user privacy through local Obsidian vault storage.
- -January 2025 – Present (Founder & Technical Lead)
- -HoverNotes is a privacy-first AI video note-taking product that turns online lectures, tutorials, and course videos into structured notes users can keep in their own Obsidian/Markdown workflows. I founded and built the product across browser-extension workflows, a Next.js web app, AI note-generation systems, payments, analytics, localization, SEO, and production operations.
+ +Next.js, React, TypeScript, browser extensions, Lexical, Python/FastAPI, PostgreSQL/Drizzle, Stripe, Razorpay, PostHog, Vercel, Cloudflare R2, and AI/LLM workflows.
+ diff --git a/_projects/2_project.md b/_projects/2_project.md index 12a57ae..9295605 100644 --- a/_projects/2_project.md +++ b/_projects/2_project.md @@ -12,7 +12,7 @@ category: workPioneered the development of AI-powered auto note generation from video content at Slid, starting as a side project before the AI boom. Built an end-to-end system that transforms video content into structured, actionable notes using advanced AI processing, achieving significant cost optimization and establishing new revenue streams.
2024 (Principal AI Engineer at Slid)
+2024 (Lead AI Engineer at Slid)
AI/ML Engineering, System Architecture, Cost Optimization, Real-time Processing, API Integration, Product Innovation, Full-stack Development
- \ No newline at end of file + diff --git a/_projects/3_project.md b/_projects/3_project.md index 63e6fe8..8a426c5 100644 --- a/_projects/3_project.md +++ b/_projects/3_project.md @@ -9,7 +9,7 @@ category: workLed technical architecture for Textutor AI Platform at GoGymi - a comprehensive AI-powered essay correction and grading system revolutionizing language education. Built sophisticated GPT-4 integration with function calling for structured educational assessment, processing 1000+ student submissions daily with 99.9% uptime.
+Led technical architecture for TexTutor, an AI-powered educational SaaS platform for Swiss teachers and students. Built full-stack product foundations across authentication, billing, multilingual product flows, AI grammar correction, collaboration, and deployment workflows.
2024 - 2025 (Technical Lead, Freelance)
@@ -20,7 +20,7 @@ category: workTechnical Leadership, AI/ML Integration, Database Design, Real-time Systems, Educational Technology, Multi-language Processing, System Architecture, Performance Optimization
-Discovered and implemented a method to extract direct YouTube video URLs, enabling server-side processing for mobile platforms. This breakthrough became the foundation for multiple features including mobile video processing and remote capture systems, supporting thousands of concurrent educational requests.
2024 (Principal AI Engineer at Slid)
+2024 (Lead AI Engineer at Slid)
Reverse Engineering, System Architecture, Python Development, API Design, Anti-Detection Systems, Performance Optimization, Production Scaling
- \ No newline at end of file + diff --git a/_projects/8_project.md b/_projects/8_project.md index 94945e6..6649b8c 100644 --- a/_projects/8_project.md +++ b/_projects/8_project.md @@ -12,7 +12,7 @@ category: workBuilt sophisticated remote screenshot capture system leveraging the YouTube URL extractor discovery to solve mobile platforms' lack of video screenshot capabilities. Implemented server-side FFmpeg processing providing pixel-perfect video frame capture for mobile users with sub-second latency.
2024 (Principal AI Engineer at Slid)
+2024 (Lead AI Engineer at Slid)
Video Processing, FFmpeg, FastAPI, Cloud Architecture, Mobile Development, Performance Optimization, System Design
- \ No newline at end of file + diff --git a/_projects/9_project.md b/_projects/9_project.md index 34bfef9..c37d7d1 100644 --- a/_projects/9_project.md +++ b/_projects/9_project.md @@ -12,7 +12,7 @@ category: workBuilt sophisticated multi-agent system for AI dubbing pipeline transcript correction using LangGraph. Developed specialized correction agents with custom reasoning tools, confidence scoring, and validation mechanisms that seamlessly integrated with existing dubbing workflows while significantly improving quality metrics.
2024 (Principal AI Engineer at Slid/Bebridge)
+2024 (Lead AI Engineer at Slid/Bebridge)
AI Agent Development, LangGraph, LangChain, Multi-Agent Systems, NLP, Production ML, System Integration, API Design
- \ No newline at end of file + diff --git a/_sass/_variables.scss b/_sass/_variables.scss index 66b3d00..6975473 100644 --- a/_sass/_variables.scss +++ b/_sass/_variables.scss @@ -3,6 +3,8 @@ * To adjust anything, simply edit the variables below and rebuild the theme. ******************************************************************************/ +@use "sass:color"; + // Colors $red-color: #FF3636 !default; $red-color-dark: #B71C1C !default; @@ -10,20 +12,20 @@ $orange-color: #F29105 !default; $blue-color: #0076df !default; $blue-color-dark: #00369f !default; $cyan-color: #2698BA !default; -$light-cyan-color: lighten($cyan-color, 25%); +$light-cyan-color: color.adjust($cyan-color, $lightness: 25%); $green-color: #00ab37 !default; $green-color-lime: #B7D12A !default; $green-color-dark: #009f06 !default; $green-color-light: #ddffdd !default; $green-color-bright: #11D68B !default; $purple-color: #B509AC !default; -$light-purple-color: lighten($purple-color, 25%); +$light-purple-color: color.adjust($purple-color, $lightness: 25%); $pink-color: #f92080 !default; $pink-color-light: #ffdddd !default; $yellow-color: #efcc00 !default; $grey-color: #828282 !default; -$grey-color-light: lighten($grey-color, 40%); +$grey-color-light: color.adjust($grey-color, $lightness: 40%); $grey-color-dark: #1C1C1D; $grey-900: #212529; @@ -36,4 +38,4 @@ $code-bg-color-light: rgba($purple-color, 0.05); $code-bg-color-dark: #2c3237 !default; // Font awesome location -$fa-font-path: "../webfonts"; \ No newline at end of file +$fa-font-path: "../webfonts"; diff --git a/_sass/font-awesome/_functions.scss b/_sass/font-awesome/_functions.scss index a17ffe8..ce759b0 100644 --- a/_sass/font-awesome/_functions.scss +++ b/_sass/font-awesome/_functions.scss @@ -3,7 +3,7 @@ // fa-content: convenience function used to set content property @function fa-content($fa-var) { - @return unquote("\"#{ $fa-var }\""); + @return "#{ $fa-var }"; } // fa-divide: Originally obtained from the Bootstrap https://github.com/twbs/bootstrap diff --git a/_sass/font-awesome/_icons.scss b/_sass/font-awesome/_icons.scss index 0f55926..e8daafd 100644 --- a/_sass/font-awesome/_icons.scss +++ b/_sass/font-awesome/_icons.scss @@ -5,6 +5,5 @@ readers do not read off random characters that represent icons */ @each $name, $icon in $fa-icons { - .#{$fa-css-prefix}-#{$name}::before { content: unquote("\"#{ $icon }\""); } + .#{$fa-css-prefix}-#{$name}::before { content: "#{ $icon }"; } } - diff --git a/_sass/font-awesome/brands.scss b/_sass/font-awesome/brands.scss index 202ed64..4278406 100644 --- a/_sass/font-awesome/brands.scss +++ b/_sass/font-awesome/brands.scss @@ -26,5 +26,5 @@ } @each $name, $icon in $fa-brand-icons { - .#{$fa-css-prefix}-#{$name}:before { content: unquote("\"#{ $icon }\""); } + .#{$fa-css-prefix}-#{$name}:before { content: "#{ $icon }"; } } diff --git a/assets/pdf/2026-06-05-Hamza-Kyamanywa-Resume-Draft.pdf b/assets/pdf/2026-06-05-Hamza-Kyamanywa-Resume-Draft.pdf new file mode 100644 index 0000000..519d1f8 Binary files /dev/null and b/assets/pdf/2026-06-05-Hamza-Kyamanywa-Resume-Draft.pdf differ diff --git a/assets/pdf/2026-06-05-Hamza-Kyamanywa-Resume-Founder-Draft.pdf b/assets/pdf/2026-06-05-Hamza-Kyamanywa-Resume-Founder-Draft.pdf new file mode 100644 index 0000000..519d1f8 Binary files /dev/null and b/assets/pdf/2026-06-05-Hamza-Kyamanywa-Resume-Founder-Draft.pdf differ diff --git a/assets/pdf/2026-06-05-Hamza-Kyamanywa-Resume-Technical-Draft.pdf b/assets/pdf/2026-06-05-Hamza-Kyamanywa-Resume-Technical-Draft.pdf new file mode 100644 index 0000000..ed41820 Binary files /dev/null and b/assets/pdf/2026-06-05-Hamza-Kyamanywa-Resume-Technical-Draft.pdf differ diff --git a/career_evidence/README.md b/career_evidence/README.md new file mode 100644 index 0000000..b9aafd1 --- /dev/null +++ b/career_evidence/README.md @@ -0,0 +1,54 @@ +# Career Evidence Bank + +This folder is the working source of truth for updating Hamza's personal site, portfolio, and resumes. + +The public site files are polished outputs. The files here are the detailed input layer: role history, project scope, dates, metrics, technical decisions, evidence links, and candidate-ready bullet material. + +## How To Use This Folder + +Before editing the site or resume: + +1. Read `index.yml` to see the role/project map. +2. Open the relevant file in `roles/`. +3. If updating a resume, review `resume_ai_extractions/` for notes on what automated systems preserved or missed. +4. Use claims marked `Verified` or backed by local repo/report evidence directly. +5. Treat `Needs confirmation` as draft material. Confirm it with Hamza or stronger evidence before publishing. +6. If a new role, project, or achievement appears, create a new file from `_TEMPLATE.md`. + +## Evidence Levels + +- `Verified` - supported by a local repo, existing resume, contribution report, public link, or explicit user statement. +- `Strong draft` - likely true based on existing materials but should be checked before prominent use. +- `Needs confirmation` - useful candidate language that requires Hamza's review before publishing. + +## Files + +- `index.yml` - machine-readable map of roles/projects and where to find source material. +- `_TEMPLATE.md` - template for adding future roles/projects. +- `linkedin/` - LinkedIn-specific profile, About, experience, skills, and Featured section drafts. +- `resume_ai_extractions/` - diagnostic notes on how AI/ATS/platform parsers interpret generated resumes. +- `roles/quest-platform.md` - Quest safari quotation platform work. +- `roles/quest-pdf-service.md` - Quest PDF generation service work. +- `roles/xai-human-data.md` - xAI Human Data / AI Tutor software engineering contract work. +- `roles/hovernotes.md` - HoverNotes founder work. +- `roles/jonjabird.md` - JonjaBird founder/operator and business automation work. +- `roles/uganda-community-korea.md` - Uganda community leadership, governance values, diaspora partnership, and sports diplomacy work. +- `roles/bebridge-slid.md` - Bebridge/Slid work split by growth period. +- `roles/gogymi-textutor.md` - GoGymi/TexTutor freelance platform work. + +## Public Artifact Targets + +Use this evidence to update: + +- Homepage: `_pages/about.md` +- CV page data: `_data/cv.yml` +- Project pages: `_projects/` +- Resume markdown: `2025-10-22-hamza-kyamanywa-resume.md` and tailored variants +- Generated PDFs under `assets/pdf/` when needed + +## Writing Rules + +- Keep resume bullets outcome-first: action, system, scale, result. +- Keep site copy human and readable; do not overload the homepage with every metric. +- Avoid duplicate or conflicting dates across files. +- Do not publish private client details, credentials, internal URLs, or unreleased financial data unless Hamza explicitly approves. diff --git a/career_evidence/_TEMPLATE.md b/career_evidence/_TEMPLATE.md new file mode 100644 index 0000000..a9890a8 --- /dev/null +++ b/career_evidence/_TEMPLATE.md @@ -0,0 +1,67 @@ +# Project Or Role Name + +## Metadata + +- Evidence status: Needs confirmation +- Public name: +- Organization/client: +- Role/title: +- Period: +- Location: +- Source repos: +- Existing public artifacts: +- Private/internal artifacts: + +## One-Line Positioning + +Short public-facing summary of what Hamza did and why it mattered. + +## Context + +What the product, company, or project does. Include audience, market, team context, and constraints. + +## Ownership + +- + +## Major Workstreams + +### Workstream Name + +- Problem: +- What Hamza built: +- Technologies: +- Result: +- Evidence: + +## Metrics And Outcomes + +- Metric: + - Status: + - Evidence: + - Public wording: + +## Technical Stack + +- Frontend: +- Backend: +- AI/ML: +- Data: +- Infrastructure: +- Integrations: + +## Resume Bullet Bank + +- + +## Site/Portfolio Angles + +- + +## Proof Links And Evidence + +- + +## Needs Confirmation + +- diff --git a/career_evidence/index.yml b/career_evidence/index.yml new file mode 100644 index 0000000..dac5aba --- /dev/null +++ b/career_evidence/index.yml @@ -0,0 +1,114 @@ +roles: + - id: xai-human-data + public_name: xAI Human Data - Software Engineering + organization: xAI / Remote Technology Services + role: Software Engineering Specialist, Human Data / AI Tutor - Software Engineering (Expert) + period: May 2026 - Present + evidence_file: roles/xai-human-data.md + source_repos: + - Private local workspace, not included in this repo. + public_artifact_targets: + - _pages/about.md + - _data/cv.yml + - resume markdown files + status: Verified from contract, user statement, and public xAI role posting; internal details confidential + + - id: quest-platform + public_name: Quest Safari Quotation Platform + organization: Quest + public_url: https://questheaven.com + role: CTO and Co-founder + period: 2026 - Present needs confirmation + evidence_file: roles/quest-platform.md + source_repos: + - private-local-workspace/quest/quest-web-app + public_artifact_targets: + - _pages/about.md + - _data/cv.yml + - _projects/ + - resume markdown files + status: Verified local repo and user-stated CTO/co-founder role, exact dates need confirmation + + - id: quest-pdf-service + public_name: Quest PDF Service + organization: Quest + role: CTO and Co-founder / PDF infrastructure owner + period: 2026 - Present needs confirmation + evidence_file: roles/quest-pdf-service.md + source_repos: + - private-local-workspace/quest/quest-pdf-service + public_artifact_targets: + - _projects/ + - resume markdown files + status: Verified local repo and user-stated CTO/co-founder role, dates need confirmation + + - id: hovernotes + public_name: HoverNotes + organization: HoverNotes + role: Technical Founder and CEO + period: Jul 2025 - Present + evidence_file: roles/hovernotes.md + source_repos: + - private-local-workspace/hovernotes/hover-notes-web-app + public_artifact_targets: + - _pages/about.md + - _data/cv.yml + - resume markdown files + status: Repo-backed working draft, user-confirmed revenue/install metrics as of 2026-06-05 + + - id: jonjabird + public_name: JonjaBird + organization: JonjaBird + role: Founder, product operator, and automation engineer + period: Mar 2026 - Present needs confirmation + evidence_file: roles/jonjabird.md + source_repos: + - private-local-workspace/inmate-letters-business + public_artifact_targets: + - _pages/about.md + - _data/cv.yml + - _projects/ + - resume markdown files + status: Verified local business workspace, public framing and metrics need confirmation + + - id: uganda-community-korea + public_name: Uganda Community in South Korea + organization: USKA / UISKA / UCISK needs confirmation + role: Vice President / executive community leadership + period: 2026 - Present needs confirmation + evidence_file: roles/uganda-community-korea.md + source_repos: + - private-local-workspace/uganda-community-in-korea + public_artifact_targets: + - _pages/about.md + - _data/cv.yml + - _projects/ + - resume markdown files + status: Verified local artifacts and user-stated VP role, public naming and dates need confirmation + + - id: bebridge-slid + public_name: Slid + organization: Bebridge Inc. + role: Full Stack Engineer to AI Engineer and Team Lead + period: Jun 2022 - Jun 2025 + evidence_file: roles/bebridge-slid.md + source_repos: [] + public_artifact_targets: + - _pages/about.md + - _data/cv.yml + - _projects/ + - resume markdown files + status: Strong draft from existing resume/site/project pages + + - id: gogymi-textutor + public_name: TexTutor + organization: GoGymi + role: Senior Full Stack Engineer / freelance technical lead + period: Jan 2025 - Sep 2025 + evidence_file: roles/gogymi-textutor.md + source_repos: [] + public_artifact_targets: + - _pages/about.md + - _data/cv.yml + - resume markdown files + status: Verified from contribution reports, final wording needs review diff --git a/career_evidence/linkedin/2026-06-05-linkedin-update-plan.md b/career_evidence/linkedin/2026-06-05-linkedin-update-plan.md new file mode 100644 index 0000000..c7828fc --- /dev/null +++ b/career_evidence/linkedin/2026-06-05-linkedin-update-plan.md @@ -0,0 +1,388 @@ +# LinkedIn Update Plan - 2026-06-05 + +## Metadata + +- Evidence status: Draft for Hamza review +- Date: 2026-06-05 +- Inputs: + - Current LinkedIn profile text pasted by user + - `career_evidence/roles/hovernotes.md` + - `career_evidence/roles/quest-platform.md` + - `career_evidence/roles/xai-human-data.md` + - `career_evidence/roles/gogymi-textutor.md` + - `career_evidence/roles/bebridge-slid.md` + - xAI platform parser extraction note +- Goal: Update LinkedIn so it presents Hamza as a founder-CTO and senior AI/full-stack product engineer, while still preserving the technical depth that automated systems and recruiters extract well. + +## Positioning Thesis + +Hamza should not look like a generic AI engineer with a long tool list. The strongest LinkedIn story is: + +> Technical founder and CTO who builds AI products for real workflows: video learning, safari quotations, and coding-model evaluation. Strong across product strategy, full-stack systems, LLM workflows, browser extensions, payments, document generation, and production reliability. + +The profile should make three things obvious: + +1. Current work is HoverNotes + Quest. +2. HoverNotes is real and monetized: `$10K+/month revenue`, around `20K Chrome/Edge installs`. +3. Quest is real CTO/co-founder work: leading a `3-person engineering/product team`, building B2B SaaS and PDF infrastructure for safari operators. + +## Headline Options + +### Recommended + +Technical Founder & CTO | Building HoverNotes ($10K+/mo, 20K installs) + Quest | AI, Full-Stack, Browser Extensions, Payments & PDF Systems + +### More Technical + +Technical Founder @ HoverNotes | CTO & Co-founder @ Quest | Senior AI / Full-Stack Engineer | Next.js, LLMs, Browser Extensions, Payments, PDF Infra + +### More Founder-Focused + +Founder-CTO building AI products for real workflows | HoverNotes, Quest | Video learning, B2B SaaS, LLM systems, payments & document infrastructure + +### Shorter / Cleaner + +Founder-CTO & Senior AI Product Engineer | HoverNotes + Quest | Full-stack, LLM systems, browser extensions, payments, PDF infrastructure + +## About Section - Recommended + +I build AI products where real workflows, technical systems, and monetization meet. + +Right now I spend most of my time on two products: + +**HoverNotes** - a privacy-first AI video note-taking product that turns online lectures, tutorials, and course videos into structured notes users can keep in their own Obsidian/Markdown workflows. I founded and built the product across the Chrome/Edge extension, Next.js web app, AI note-generation flows, payments, analytics, localization, SEO, and production operations. As of June 2026, HoverNotes is doing over $10K/month in revenue with around 20K Chrome/Edge installs. + +**Quest** - a safari quotation platform for tour operators. I co-founded Quest as CTO and lead a 3-person engineering/product team building the platform: operator dashboards, admin tools, travel requests, itinerary planning, pricing, branded PDF generation, quote email delivery, content libraries, billing, and release workflows. + +I also contract with xAI Human Data on software-engineering model evaluation, reviewing AI-generated code across languages and domains to help improve coding-model quality. + +Before this, I was a Lead AI Engineer at Bebridge/Slid, a 50K+ MAU EdTech platform, where I took on principal-level AI engineering work: setting technical direction, solving the hardest AI/product problems, and shipping LLM-powered auto-notes, AI chat, semantic search, transcript correction, mobile AI workflows, payment/auth systems, real-time transcription, and multi-format exports. + +Technically, I work across TypeScript, React, Next.js, Python/FastAPI, PostgreSQL, Drizzle, browser extensions, LangChain/LangGraph, OpenAI/Whisper, Stripe/Razorpay, Cloudflare R2, Puppeteer/Handlebars PDF generation, Vercel, AWS, Docker, PostHog, and Sentry. + +I’m strongest when the work requires both product judgment and engineering depth: understanding a messy workflow, designing the business loop, building the full stack, instrumenting behavior, and making the system reliable enough for users to trust. + +## About Section - More Technical + +I’m a technical founder, CTO, and senior AI/full-stack engineer building production AI products across education, productivity, travel, and coding-model evaluation. + +My current focus is HoverNotes and Quest. + +At **HoverNotes**, I founded and built a privacy-first AI video note-taking SaaS with over $10K/month revenue and around 20K Chrome/Edge installs. The product spans a Chrome/Edge extension, Next.js web app, local-first Obsidian workflows, AI note-generation pipelines, regional payments, analytics, SEO/i18n, and production reliability. + +At **Quest**, I’m CTO and co-founder, leading a 3-person team building a Next.js safari quotation platform for Uganda-based tour operators. The platform includes operator/admin workflows, structured destination/activity/hotel/transport data, quote pricing, PDF preview/export, Resend quote email delivery, Stripe billing, Cloudflare R2 storage, and a TypeScript/Express + Puppeteer PDF service. + +I also work with xAI Human Data as a software-engineering specialist / AI Tutor, evaluating AI-generated code for model training and benchmarking across Python, TypeScript, Java, Go, Rust, C/C++, databases, distributed systems, AI/ML, security, and performance domains. + +Earlier at Bebridge/Slid, I grew from early full-stack engineer to AI team lead, shipping LLM-powered auto-notes, AI chat, semantic search, transcript correction, React Native AI workflows, real-time transcription, cross-platform auth/payments, and export systems on a 50K+ MAU EdTech platform. + +Core stack: TypeScript, React, Next.js, Python/FastAPI, PostgreSQL, Supabase, Drizzle, Node/Express, LangChain, LangGraph, OpenAI, Whisper, browser extensions, Lexical, Stripe, Razorpay, Puppeteer, Handlebars, Cloudflare R2, Vercel, AWS, Docker, PostHog, Sentry. + +## Experience Updates + +### HoverNotes + +Recommended LinkedIn title: + +Technical Founder & CEO + +Employment type: + +Full-time + +Location: + +Remote / Seoul + +Dates: + +Jul 2025 - Present + +Description: + +Founded and built HoverNotes, a privacy-first AI video note-taking SaaS that turns online lectures, tutorials, and course videos into structured notes users can keep in their own Obsidian/Markdown workflows. + +As of June 2026, HoverNotes is doing over $10K/month in revenue with around 20K installs across Chrome and Edge. + +Key work: + +- Built the product across Chrome/Edge extension workflows, a Next.js web app, local-first note-taking flows, AI note generation, payments, analytics, SEO/i18n, and production operations. +- Designed a multimodal video-in, notes-out pipeline where the extension captures realtime video learning context, the web app deduplicates/ranks visual frames, and the backend generates structured notes from transcript plus selected visual context. +- Built Obsidian vault saving, screenshot handling, note history, reconnect/recovery paths, and guardrails against destructive auto-save behavior. +- Implemented regional monetization across Stripe for global card payments, China one-time passes via WeChat Pay/Alipay, and India payments via Razorpay/UPI. +- Used analytics to diagnose AI Notes activation, paywall behavior, platform-specific drop-off, media payload size, warnings, and active-session health. + +Suggested skills: + +Next.js, TypeScript, React, Chrome Extensions, Browser Extensions, OpenAI API, LangGraph, FastAPI, Stripe, Razorpay, PostHog, Obsidian, Product Analytics, SEO + +### Quest + +Recommended LinkedIn title: + +CTO & Co-founder + +Employment type: + +Full-time or Self-employed needs confirmation + +Location: + +Remote + +Dates: + +Jan 2026 - Present if you want to match the xAI parser inference; otherwise 2026 - Present if LinkedIn allows year-only. + +Description: + +Co-founded Quest as CTO and lead a 3-person engineering/product team building a safari quotation platform for Uganda-based tour operators. + +Quest turns operator quoting from spreadsheets, scattered assets, and manual documents into a structured workflow for travel requests, itineraries, quote pricing, branded PDFs, quote email delivery, billing, analytics, and content libraries. + +Key work: + +- Lead product and platform architecture across operator dashboards, Quest admin tools, travel requests, tourists, content libraries, quote building, analytics, billing, company settings, and role-aware access. +- Built structured domain workflows for destinations, activities, hotels/lodges, room rates, transport, geography, company content, vehicles, staff, reviews, and per-company settings. +- Designed quote economics for real operator constraints, including USD/UGX exchange rates, transport fuel conversion, room pricing, service fees, and quote-level cost summaries. +- Built the document delivery system as a core product surface: branded A4 safari PDFs, page previews, edit overrides, Cloudflare R2 storage, Resend quote emails, merge tags, sender flows, and PDF attachments. +- Designed the Free + Pro launch model with company-level Stripe subscriptions, full-access trials, billing settings, plan badges, and output-level gates for clean PDFs, email sending, and branding. +- Established Vercel QA/release workflows and deployed PDF infrastructure with Docker/Hetzner. + +Suggested skills: + +Next.js, TypeScript, PostgreSQL, Supabase, Drizzle ORM, NextAuth, Stripe, Puppeteer, Handlebars, Cloudflare R2, Resend, Product Management, Technical Leadership + +### xAI / Remote Technology Services + +Recommended LinkedIn title: + +Software Engineering Specialist, Human Data / AI Tutor - Software Engineering + +Employment type: + +Contract + +Location: + +Remote + +Dates: + +May 2026 - Present + +Description: + +Contracted with xAI's Human Data software-engineering track to create and review high-quality coding data for training, benchmarking, and improving large language models. + +Key work: + +- Evaluate AI-generated code across correctness, maintainability, performance, security, test coverage, task alignment, and production plausibility. +- Rank model outputs using evidence from prompts, diffs, traces, responses, tests, and container checks. +- Write concise, evidence-grounded justifications for model training and benchmarking workflows. +- Review coding tasks across Python, TypeScript/JavaScript, Java, Go, Rust, C/C++, databases, distributed systems, AI/ML, security, and performance domains. + +Suggested skills: + +Software Engineering, Code Review, Python, TypeScript, JavaScript, Java, Go, Rust, C++, AI Evaluation, LLM Evaluation, Model Evaluation + +Public caution: + +Keep this high-level. Do not publish internal task names, prompts, platform names beyond what is public/approved, private links, or workflow details that belong inside the contract environment. + +### GoGymi / TexTutor + +Recommended LinkedIn title: + +Senior Full Stack Engineer / Freelance Technical Lead + +Employment type: + +Freelance + +Location: + +Remote, Switzerland + +Dates: + +Jan 2025 - Sep 2025 + +Description: + +Led full-stack development of TexTutor, an AI-powered EdTech SaaS platform for Swiss teachers and students, from architecture through billing, AI correction, collaboration, and deployment workflows. + +Key work: + +- Built the product foundation with Next.js 15, React 19, TypeScript, PostgreSQL, Drizzle ORM, Supabase, Tailwind CSS, and Docker. +- Implemented Stripe subscriptions with webhooks, trials, usage tracking, plan enforcement, billing dashboards, and multilingual billing emails. +- Built multilingual product flows across English, German, French, Spanish, and Italian. +- Integrated AI grammar-correction workflows with formatting preservation, preset/rule management, usage validation, and educational feedback loops. +- Built rich assignment and feedback experiences with Lexical, LiveBlocks collaboration, comments, notifications, sharing, and PDF export. +- Contributed 550+ commits across frontend, backend, billing, authentication, AI workflows, and deployment. + +Suggested skills: + +Next.js, React, TypeScript, PostgreSQL, Drizzle ORM, Stripe, Supabase, Lexical, LiveBlocks, Docker, EdTech + +### Bebridge / Slid - Lead AI Engineer / Team Lead + +Recommended LinkedIn title: + +Lead AI Engineer / Team Lead + +Dates: + +Jan 2024 - Jun 2025 + +Description: + +Led AI product development for Slid, a 50K+ MAU EdTech platform, taking on principal-level AI engineering work across web, mobile, and backend systems. + +Key work: + +- Shipped LLM-powered auto-notes, AI chat, semantic search, transcript correction, and mobile AI workflows. +- Set technical direction for hard AI/product problems across video learning, mobile constraints, AI quality, and production cost. +- Built LLM/Whisper-based auto-notes from video lectures and contributed to measurable engagement and retention gains. +- Spearheaded React Native mobile AI features backed by AWS Lambda and FFmpeg video processing. +- Designed LangGraph multi-agent transcript-correction workflows with custom reasoning tools. +- Optimized AI processing costs through buffering, caching, provider strategy, and production pipeline improvements. + +Suggested skills: + +LLMs, LangGraph, Whisper, React Native, AWS Lambda, FFmpeg, Semantic Search, AI Product Development + +### Bebridge / Slid - Senior Software Engineer / Full Stack Engineer + +Recommended LinkedIn title: + +Senior Software Engineer / Full Stack Engineer + +Dates: + +Jun 2022 - Dec 2023 + +Description: + +Joined Slid as an early engineer and built core systems across web, Chrome extension, desktop, mobile, payments, authentication, transcription, and AI learning features. + +Key work: + +- Built early RAG and conversational AI experiences for learning notes using LangChain, Pinecone, embeddings, and streaming responses. +- Pioneered real-time transcription using WebRTC, Socket.io, and cloud speech-to-text providers. +- Built cross-platform authentication and payments across web, extension, mobile, and desktop, including AWS Cognito, Stripe, and Payple. +- Shipped multi-format export workflows to Word, PDF, Markdown, and Notion, reducing export-related support burden. +- Earned Bebridge Hero Awards in 2022 and 2023 and was promoted from Level 2 to Level 4 in 18 months. + +Suggested skills: + +WebRTC, Socket.io, LangChain, Pinecone, AWS Cognito, Stripe, Payple, Chrome Extensions, Full-Stack Development + +## Volunteer / Leadership Section + +### Ugandan Community in South Korea + +Recommended LinkedIn title: + +Vice President + +Organization: + +Ugandan Community in South Korea + +Dates: + +2026 - Present + +Description: + +Serve as Vice President of the Ugandan Community in South Korea, supporting governance, embassy-facing communication, partnership strategy, and community engagement. + +Key work: + +- Helped formalize leadership values and governance materials around respect, servant leadership, integrity, transparency, and accountability. +- Supported embassy-facing and institutional partnership materials for cultural diplomacy, sports/community programs, student/professional opportunities, and Uganda-Korea relations. +- Supported community communication and reporting around diaspora activities, including sports diplomacy and KAFCON 2026. + +Suggested skills: + +Leadership, Community Engagement, Public Speaking, Governance, Stakeholder Communication, Partnership Development + +## Featured Section Ideas + +Add 3-5 featured items if possible: + +1. HoverNotes website or Chrome Web Store listing. +2. Quest website or a public product/demo page if available. +3. Resume PDF: `2026-06-05-Hamza-Kyamanywa-Resume-Technical-Draft.pdf`. +4. Selected open-source contribution or GitHub profile. +5. Springer Nature paper or research publication. + +## Skills To Prioritize + +LinkedIn lets you pin top skills. Recommended top skills: + +1. AI Product Development +2. Next.js +3. TypeScript +4. LLM Applications +5. Full-Stack Development +6. Technical Leadership +7. Browser Extensions +8. Stripe +9. PostgreSQL +10. Product Analytics + +Additional skills worth keeping: + +- React +- Python +- FastAPI +- LangChain +- LangGraph +- OpenAI API +- Whisper +- RAG +- Vector Search +- Supabase +- Drizzle ORM +- Docker +- Vercel +- AWS Lambda +- Cloudflare R2 +- Puppeteer +- Handlebars +- WebRTC +- Socket.io +- React Native +- Lexical +- LiveBlocks +- PostHog +- Sentry +- Razorpay +- PDF generation +- SEO +- i18n +- Korean + +## Profile Cleanup Notes + +- Remove or reduce emoji-heavy section labels in older experience entries. They are memorable, but the profile now needs a cleaner senior/founder tone. +- Add Quest and xAI as current roles so LinkedIn reflects current reality. +- Update HoverNotes location from `United States - Hybrid` if that is no longer the desired public framing. +- Avoid saying `zero cloud data storage` unless it is strictly true. Safer phrasing: `local-first workflows` and `privacy-first note ownership`. +- Replace `45+ language localization` if current supported production locales are closer to 10. Safer phrasing: `localized UI/content and multilingual AI note workflows`. +- Keep Slid technical depth, but compress it slightly so HoverNotes + Quest feel like the current center of gravity. +- Consider adding `Vice President, Ugandan Community in South Korea` under Volunteer Experience rather than Experience, unless you want civic leadership to be a major professional signal. + +## Suggested Update Order + +1. Update headline. +2. Replace About section. +3. Add Quest current role. +4. Add xAI contract current role. +5. Refresh HoverNotes description and metrics. +6. Compress Slid and TexTutor entries. +7. Add community Vice President role under Volunteer Experience. +8. Update Featured links. +9. Pin top skills. diff --git a/career_evidence/linkedin/2026-06-05-new-about-section.md b/career_evidence/linkedin/2026-06-05-new-about-section.md new file mode 100644 index 0000000..a5af3e1 --- /dev/null +++ b/career_evidence/linkedin/2026-06-05-new-about-section.md @@ -0,0 +1,114 @@ +# New LinkedIn About Section - 2026-06-05 + +## Metadata + +- Evidence status: Draft for Hamza review +- Date: 2026-06-05 +- Purpose: New LinkedIn About section drafted after preserving the previous About snapshot. +- Positioning: Technical founder, CTO, and senior AI/full-stack product engineer focused on HoverNotes, Quest, and xAI contract work. + +## Recommended About Section + +I’m a technical founder, CTO, and AI/full-stack engineer building products where real workflows, technical systems, and monetization meet. + +Right now I spend most of my time on two products: + +**HoverNotes** - a privacy-first AI video note-taking product that turns online lectures, tutorials, and course videos into structured notes users can keep in their own Obsidian/Markdown workflows. I founded and built the product across the Chrome/Edge extension, Next.js web app, realtime video-in notes-out AI workflows, payments, analytics, localization, SEO, and production operations. As of June 2026, HoverNotes is doing over $10K/month in revenue with around 20K installs across Chrome and Edge. + +**Quest** - a safari quotation platform for tour operators. I co-founded Quest as CTO and lead a 3-person engineering/product team building the platform: operator dashboards, admin tools, travel requests, itinerary planning, pricing, branded PDF generation, quote email delivery, content libraries, billing, and release workflows. + +I also contract with xAI Human Data on software-engineering model evaluation, reviewing AI-generated code across languages and domains to help improve coding-model quality. + +Before this, I was a Lead AI Engineer at Bebridge/Slid, a 50K+ MAU EdTech platform, where I took on principal-level AI engineering work: setting technical direction, solving the hardest AI/product problems, and shipping LLM-powered auto-notes, AI chat, semantic search, transcript correction, mobile AI workflows, payment/auth systems, real-time transcription, and multi-format exports. + +My engineering philosophy is simple: + +- Put users at the center, especially when the technical system is complex. +- Move fast, but instrument behavior so decisions are grounded in real usage. +- Build the whole loop: product, backend, frontend, payments, analytics, reliability, and iteration. +- Stay calm under ambiguity and turn messy workflows into software people can trust. + +Technically, I work across TypeScript, React, Next.js, Python/FastAPI, PostgreSQL, Drizzle, browser extensions, LangChain/LangGraph, OpenAI/Whisper, Stripe/Razorpay, Cloudflare R2, Puppeteer/Handlebars PDF generation, Vercel, AWS, Docker, PostHog, and Sentry. + +I’m especially interested in privacy-first AI products, AI learning tools, coding agents/model evaluation, workflow automation, and products that require both product judgment and deep engineering execution. + +## Improved Hybrid Version + +I’m a technical founder, CTO, and AI/full-stack engineer with 4+ years of hands-on experience building AI-powered education, productivity, travel, and workflow products. + +My work sits at the intersection of product judgment and deep engineering execution: turning messy real-world workflows into software people can trust, use, and pay for. + +Right now I’m focused on two products: + +**HoverNotes** - a privacy-first AI video note-taking product that turns online lectures, tutorials, and course videos into structured notes users can keep in their own Obsidian/Markdown workflows. I founded and built the product across the Chrome/Edge extension, Next.js web app, realtime video-in notes-out AI workflows, payments, analytics, localization, SEO, and production operations. As of June 2026, HoverNotes is doing over $10K/month in revenue with around 20K installs across Chrome and Edge. + +**Quest** - a safari quotation platform for tour operators. I co-founded Quest as CTO and lead a 3-person engineering/product team building operator dashboards, admin tools, travel requests, itinerary planning, pricing, branded PDF generation, quote email delivery, content libraries, billing, and release workflows. + +I also contract with xAI Human Data on software-engineering model evaluation, reviewing AI-generated code across languages and domains to help improve coding-model quality. + +Before this, I was a Lead AI Engineer at Bebridge/Slid, a 50K+ MAU EdTech platform, where I took on principal-level AI engineering work: setting technical direction, solving the hardest AI/product problems, and shipping LLM-powered auto-notes, AI chat, semantic search, transcript correction, mobile AI workflows, payment/auth systems, real-time transcription, and multi-format exports. + +A few things I’m proud of beyond the role titles: + +- Co-authored a Springer Nature Computer Science paper on responsible AI and interpretable machine learning. +- Contributed to developer tools and AI/open-source ecosystems including LangChain, Vercel AI SDK, Editor.js, Redux Toolkit, MemGPT, Pytube, FastAPI, and related projects. +- Published technical answers and references used by thousands of developers working with WebRTC, speech-to-text, and AI application patterns. +- Serve as Vice President of the Ugandan Community in South Korea, supporting governance, embassy-facing communication, partnership strategy, and community engagement. + +My engineering philosophy is simple: + +- Put users at the center, especially when the technical system is complex. +- Move fast, but instrument behavior so decisions are grounded in real usage. +- Build the whole loop: product, backend, frontend, payments, analytics, reliability, and iteration. +- Stay calm under ambiguity and turn big ideas into production systems. + +Core stack: TypeScript, React, Next.js, Python/FastAPI, PostgreSQL, Drizzle, browser extensions, LangChain/LangGraph, OpenAI/Whisper, Stripe/Razorpay, Cloudflare R2, Puppeteer/Handlebars PDF generation, Vercel, AWS, Docker, PostHog, and Sentry. + +I’m open to senior/principal engineering conversations, technical advisory, AI product collaborations, and work around privacy-first AI, AI learning tools, coding agents/model evaluation, and workflow automation. + +## Improved Hybrid Version - Under 2,600 Characters + +I’m a technical founder, CTO, and AI/full-stack engineer with 4+ years of hands-on experience building AI-powered education, productivity, travel, and workflow products. + +My work sits at the intersection of product judgment and deep engineering execution: turning messy real-world workflows into software people can trust, use, and pay for. + +Right now I’m focused on two products: + +**HoverNotes** - a privacy-first AI video note-taking product that turns online lectures, tutorials, and course videos into structured notes users can keep in their own Obsidian/Markdown workflows. I founded and built the product across the Chrome/Edge extension, Next.js web app, realtime video-in notes-out AI workflows, payments, analytics, localization, SEO, and production operations. As of June 2026, HoverNotes is doing over $10K/month in revenue with around 20K installs across Chrome and Edge. + +**Quest** - a safari quotation platform for tour operators. I co-founded Quest as CTO and lead a 3-person engineering/product team building operator dashboards, admin tools, travel requests, itinerary planning, pricing, branded PDF generation, quote email delivery, content libraries, billing, and release workflows. + +I also contract with xAI Human Data on software-engineering model evaluation, reviewing AI-generated code across languages and domains to help improve coding-model quality. + +Previously, I was a Lead AI Engineer at Bebridge/Slid, a 50K+ MAU EdTech platform, doing principal-level AI engineering work: setting technical direction, solving hard AI/product problems, and shipping LLM-powered auto-notes, AI chat, semantic search, transcript correction, mobile AI workflows, real-time transcription, and cross-platform product systems. + +Beyond role titles, I’ve co-authored a Springer Nature Computer Science paper on responsible AI, contributed to developer tools and AI/open-source ecosystems, published technical answers used by thousands of developers, and serve as Vice President of the Ugandan Community in South Korea. + +My engineering philosophy: put users at the center, move fast but measure real behavior, build the whole loop from product to reliability, and stay calm under ambiguity. + +Core stack: TypeScript, React, Next.js, Python/FastAPI, PostgreSQL, Drizzle, browser extensions, LangChain/LangGraph, OpenAI/Whisper, Stripe/Razorpay, Cloudflare R2, Puppeteer, Vercel, AWS, Docker, PostHog, and Sentry. + +Open to senior/principal engineering conversations, technical advisory, AI product collaborations, privacy-first AI, coding agents/model evaluation, and workflow automation. + +## Shorter Version + +I’m a technical founder, CTO, and AI/full-stack engineer building products where real workflows, technical systems, and monetization meet. + +Currently, I’m focused on **HoverNotes** and **Quest**. + +At HoverNotes, I founded and built a privacy-first AI video note-taking product with over $10K/month revenue and around 20K Chrome/Edge installs. The product turns online lectures, tutorials, and course videos into structured notes users can keep in their own Obsidian/Markdown workflows. + +At Quest, I’m CTO and co-founder, leading a 3-person engineering/product team building a safari quotation platform for tour operators: requests, itineraries, pricing, branded PDFs, quote emails, content libraries, billing, and release workflows. + +I also contract with xAI Human Data on software-engineering model evaluation, reviewing AI-generated code across languages and domains to improve coding-model quality. + +Previously, I was a Lead AI Engineer at Bebridge/Slid, a 50K+ MAU EdTech platform, doing principal-level AI engineering work across LLM-powered auto-notes, AI chat, semantic search, transcript correction, mobile AI workflows, real-time transcription, and cross-platform product systems. + +Core stack: TypeScript, React, Next.js, Python/FastAPI, PostgreSQL, Drizzle, browser extensions, LangChain/LangGraph, OpenAI/Whisper, Stripe/Razorpay, Puppeteer/Handlebars, Cloudflare R2, Vercel, AWS, Docker, PostHog, and Sentry. + +## Notes + +- This version intentionally avoids older unverified claims such as `45+ language support`, `1000+ daily submissions`, `99.9% uptime`, `100,000+ users`, `186+ repositories`, and `535+ stars`. +- It keeps the old About section's energy around engineering philosophy, but updates the public center of gravity to HoverNotes, Quest, and xAI. +- It should fit LinkedIn's About section comfortably. +- The improved hybrid version brings back the old section's proof/recognition layer: publication, open source, developer-community impact, civic leadership, and open-to signal. diff --git a/career_evidence/linkedin/2026-06-05-previous-about-section.md b/career_evidence/linkedin/2026-06-05-previous-about-section.md new file mode 100644 index 0000000..05b0bac --- /dev/null +++ b/career_evidence/linkedin/2026-06-05-previous-about-section.md @@ -0,0 +1,54 @@ +# Previous LinkedIn About Section Snapshot - 2026-06-05 + +## Metadata + +- Evidence status: User-provided current/previous LinkedIn About section +- Date captured: 2026-06-05 +- Purpose: Preserve the previous LinkedIn About section before rewriting it, including tone, positioning, claims, and older metrics that may need review before reuse. + +## Raw Previous About Section + +I am a Technical Founder and AI Engineer specializing in full-stack + AI product development with 4+ years of hands-on experience building revolutionary educational technology. Currently founding HoverNotes.io while freelancing as AI Lead at GoGymi, where I architect AI-powered educational platforms processing 1000+ daily submissions. + +🚀 Current Focus: Dual Innovation Track (2025) + +HoverNotes (Technical Founder): Building privacy-first AI video learning platform with universal video integration, local Obsidian storage, and 45+ language support. + +GoGymi AI/Tech Lead: Architecting Textutor AI platform for essay correction and grading, featuring LLM integration, multi-language processing, and real-time collaborative editing with 99.9% uptime. + +🏆 Previous R&D Leadership at Slid + +Pioneered AI-powered educational technology as Lead AI Engineer, leading development of revolutionary features: Auto Notes System (industry-first AI note generation from video content with 60% cost optimization), Smart Live Text (25% premium subscription increase), AI Sliddy chatbot (significant engagement boost), Universal Export system (75% reduction in support requests), and multi-agent transcript correction systems using LangGraph for AI dubbing pipelines. + +🏆 Proven Track Record + +- Research Publication: Co-authored paper in SN Computer Science (Springer) on responsible AI and explainable machine learning +- Industry Impact: Built features serving 100,000+ users with measurable business outcomes +- Open Source Leadership: 186+ repositories, 535+ stars, active contributor to cutting-edge AI frameworks like: +- Community Recognition: Stack Overflow solutions helping 5,000+ developers, Moderator election invitation + +💡 Engineering Philosophy & Process + +- User-Centered: Put users at the center of all decisions with rapid prototyping and iteration +- Execution Excellence: Excel at transforming big ideas into production-ready features within constraints +- Data-Driven: Trust instinct but rely on data for critical decisions, focusing on delivering value fast +- Continuous Learning: Never settle, stay calm under adversity, make high-quality user-centered decisions + +🔧 Technical Expertise + +Core Technologies: TypeScript, Python, Node.js, React, Next.js, FastAPI, AWS, Docker, PostgreSQL + +AI/ML Stack: LangChain, Vector Databases, LLM integration, Real-time processing, Multi-modal AI + +Domain Mastery: Full-stack web development, Machine Learning, Cloud Computing, DevOps, LLMOps + +Open to: Senior engineering roles, R&D positions, technical advisory, and collaboration on privacy-first AI initiatives. + +## Notes For Rewrite + +- The old version has strong energy and a clear technical/founder voice, but several claims should be reviewed before reuse. +- `GoGymi AI Lead` is now historical rather than current. +- HoverNotes should now include updated scale: over `$10K/month revenue` and around `20K Chrome/Edge installs`. +- Quest and xAI are missing because this section predates the current role mix. +- `45+ language support`, `1000+ daily submissions`, `99.9% uptime`, `100,000+ users`, `186+ repositories`, and `535+ stars` should be verified before using in the new LinkedIn About section. +- The engineering philosophy section is useful and should probably be preserved in a shorter, more mature form. diff --git a/career_evidence/linkedin/README.md b/career_evidence/linkedin/README.md new file mode 100644 index 0000000..2a19c15 --- /dev/null +++ b/career_evidence/linkedin/README.md @@ -0,0 +1,7 @@ +# LinkedIn Drafts + +This folder contains LinkedIn-specific positioning, headline, About, Experience, Featured, and skills drafts. + +LinkedIn copy can be warmer and more narrative than resume copy, but it should still use the same evidence rules as the rest of `career_evidence/`: do not invent metrics, dates, titles, or confidential details. + +Preserve older LinkedIn copy snapshots here before replacing public profile text. Older copy is useful for tone, positioning, and forgotten claims, but metrics from older snapshots should be verified before reuse. diff --git a/career_evidence/resume_ai_extractions/2026-06-05-xai-employee-platform-technical-resume.md b/career_evidence/resume_ai_extractions/2026-06-05-xai-employee-platform-technical-resume.md new file mode 100644 index 0000000..50462ed --- /dev/null +++ b/career_evidence/resume_ai_extractions/2026-06-05-xai-employee-platform-technical-resume.md @@ -0,0 +1,173 @@ +# xAI Employee Platform Resume Extraction - Technical Resume + +## Metadata + +- Evidence status: User-provided platform extraction snapshot +- Date observed: 2026-06-05 +- Source resume: `site-repo/assets/pdf/2026-06-05-Hamza-Kyamanywa-Resume-Technical-Draft.pdf` +- Platform/context: xAI / TeachX employee platform resume parser +- Purpose: Track what AI/ATS-style systems extract from Hamza's resume so future resume edits can be tested against actual parser behavior. + +## Extracted Work Experience + +### Technical Founder & CEO + +- Company: HoverNotes +- Location: Remote / Seoul +- Dates: Jul 2025 - Present +- Current: Yes +- Platform summary: + - Founded AI video note-taking SaaS with $10K+/month revenue and 20K installs. + - Built Chrome/Edge extension, Next.js app, Obsidian workflows, Stripe/Razorpay payments, and multimodal AI pipelines. + +### CTO & Co-founder + +- Company: Quest +- Location: Remote +- Dates extracted: Jan 2026 - Present +- Current: Yes +- Platform summary: + - Led 3-person team building Next.js safari quotation platform and PDF infrastructure. + - Implemented Stripe subscriptions, PostgreSQL/Supabase, Drizzle ORM, and Puppeteer PDF generation. +- Parser note: + - Resume source says `2026 - Present`; the platform inferred `Jan 2026 - Present`. + +### Software Engineering Specialist, Human Data / AI Tutor + +- Company: xAI +- Location: Remote +- Dates: May 2026 - Present +- Current: Yes +- Platform summary: + - Evaluated and reviewed AI-generated code for model training across Python, TypeScript, Java, Go, Rust, C/C++, databases, and security domains. + +### Senior Full Stack Engineer / Freelance Technical Lead + +- Company: GoGymi / TexTutor +- Location: Remote, Switzerland +- Dates: Jan 2025 - Sep 2025 +- Platform summary: + - Built AI EdTech SaaS with Next.js 15, React 19, Stripe subscriptions, multilingual support, AI grammar correction, and Lexical/LiveBlocks collaboration features. + +### AI Engineer & Team Lead + +- Company: Bebridge Inc. / Slid +- Location: Seoul, South Korea +- Dates: Jan 2024 - Jun 2025 +- Platform summary: + - Led LLM-powered auto-notes, AI chat, semantic search, and React Native mobile AI features. + - Optimized costs with LangGraph, Whisper, AWS Lambda, and caching strategies. + +### Senior Software Engineer / Full Stack Engineer + +- Company: Bebridge Inc. / Slid +- Location: Seoul, South Korea +- Dates: Jun 2022 - Dec 2023 +- Platform summary: + - Built core systems including RAG/conversational AI with LangChain/Pinecone, real-time transcription, cross-platform auth/payments, and multi-format exports. + - Promoted from L2 to L4. + +## Extracted Education + +### Korea University + +- Degree: B.S. in Electrical and Electronics Engineering +- Location: Seoul, South Korea +- Dates: Mar 2019 - Feb 2023 +- Platform summary: + - GPA 4.02/4.5. + - Magna Cum Laude. + - Global Korea Scholarship and Ugandan Government Scholarship recipient. + - Co-authored paper on interpretable music genre classification, Springer Nature. + +## Extracted Skills + +- OpenAI API +- Whisper +- LangChain +- LangGraph +- RAG +- vector search +- prompt engineering +- multimodal AI workflows +- Next.js +- React +- TypeScript +- React Native +- Tailwind CSS +- Lexical +- Chrome/Edge extensions +- Obsidian +- local-first workflows +- Python +- FastAPI +- Node.js +- Express +- PostgreSQL +- Supabase +- Drizzle ORM +- MongoDB +- Redis +- AWS Lambda +- S3 +- EventBridge +- Docker +- Vercel +- Cloudflare R2 +- Hetzner +- CI/CD +- PostHog +- Sentry +- Stripe +- Razorpay +- webhooks +- subscriptions +- trials +- pricing +- credits +- PDF/document generation +- Korean (TOPIK 5) +- English +- Luganda +- WebRTC +- Socket.io +- AWS Cognito +- Payple +- Pinecone +- embeddings +- FFmpeg +- Handlebars +- Puppeteer +- NextAuth v5 +- LiveBlocks + +## What The Parser Preserved Well + +- Current focus on HoverNotes, Quest, and xAI was extracted correctly. +- HoverNotes scale was preserved as `$10K+/month revenue` and `20K installs`. +- Quest leadership was preserved as a `3-person team`. +- The technical stack came through strongly, especially AI/LLM, Next.js/React/TypeScript, payments, PDF generation, and infrastructure. +- Slid's 50K+ MAU context was preserved in the employer line. +- Promotion from L2 to L4 survived extraction. +- Education honors and scholarships survived extraction. + +## What The Parser Simplified Or Lost + +- HoverNotes regional payment nuance was collapsed into `Stripe/Razorpay payments`; WeChat Pay/Alipay and China/India/global segmentation did not appear in the summarized role text, though the skills list retained Stripe and Razorpay. +- Quest's safari-domain workflow depth was compressed to `quotation platform and PDF infrastructure`. +- xAI title was shortened to `Software Engineering Specialist, Human Data / AI Tutor`, dropping the full `Software Engineering (Expert)` wording. +- Open source/community details were mostly converted into skills, not experience. +- Ugandan Community Vice President role did not appear as a separate work/leadership entry. +- Quest dates were inferred as `Jan 2026 - Present` from a source that only said `2026 - Present`. + +## Resume Implications + +- If a platform parser is the target, explicit skills lists work: the parser extracted a broad and accurate technical skill set. +- If a specific regional payment story matters, phrase it directly in the role bullet rather than relying on the skills section. +- If community leadership matters, it likely needs its own clearly labeled section or role-like entry; otherwise parsers may ignore it. +- If exact Quest start month matters, add the month explicitly in the source resume to avoid parser inference. +- The technical resume successfully communicates an AI/full-stack/payments/PDF/platform profile to automated systems. + +## Sensitive/Private Notes + +- The platform account details and external work email from the user-provided screenshot should not be published on the public site or resume unless explicitly approved. diff --git a/career_evidence/resume_ai_extractions/README.md b/career_evidence/resume_ai_extractions/README.md new file mode 100644 index 0000000..2861b0b --- /dev/null +++ b/career_evidence/resume_ai_extractions/README.md @@ -0,0 +1,5 @@ +# Resume AI Extractions + +This folder records how resume parsers, ATS systems, contractor platforms, and other AI extraction tools interpret Hamza's resumes. + +Use these notes when editing resumes to understand what automated systems preserve, simplify, or miss. These files are not public-facing copy; they are diagnostic evidence for future resume/site agents. diff --git a/career_evidence/roles/bebridge-slid.md b/career_evidence/roles/bebridge-slid.md new file mode 100644 index 0000000..64cd18c --- /dev/null +++ b/career_evidence/roles/bebridge-slid.md @@ -0,0 +1,137 @@ +# Bebridge Inc. / Slid + +## Metadata + +- Evidence status: Strong draft from existing resume/site/project pages +- Public name: Slid +- Organization/client: Bebridge Inc. +- Role/title progression: Full Stack Engineer (Level 2), Senior Software Engineer (Level 3), AI Engineer and Team Lead (Level 4) +- Public title preference for LinkedIn: Lead AI Engineer / Team Lead. User noted on 2026-06-05 that principal-level AI engineering scope is accurate because he set direction and solved the hardest problems, but `Principal AI Engineer` should be treated as scope language unless confirmed as an official title. +- Period: Jun 2022 - Jun 2025 +- Location: Seoul, South Korea +- Source repos: Private/not currently linked +- Existing public artifacts: `_pages/about.md`, `_data/cv.yml`, `_projects/`, resume markdown files + +## One-Line Positioning + +Joined Bebridge as an early engineer and grew into a Lead AI Engineer / Team Lead, building cross-platform EdTech and AI features for Slid's 50K+ MAU learning platform. + +## Context + +Slid is an EdTech platform for learning from online videos. Existing site/resume materials position Hamza as an early engineer who helped scale product surfaces across web, Chrome Extension, desktop, mobile, and AI-powered learning features. + +## Role Periods + +### Full Stack Engineer (Level 2) + +- Period: Jun 2022 - May 2023 +- Focus: Web, extension, desktop, mobile foundations; real-time transcription; authentication; payments; performance. +- Recognition: Hero Award 2022; promoted to Level 3. + +### Senior Software Engineer (Level 3) - R&D And AI Innovation + +- Period: Jun 2023 - Dec 2023 +- Focus: RAG, AI chat, semantic search, video summarization, export systems, performance, open source. +- Recognition: Hero Award 2023; promoted to Level 4. + +### AI Engineer And Team Lead (Level 4) + +- Period: Jan 2024 - Jun 2025 +- Focus: AI product leadership, mobile app, auto-notes, server-side video processing, cost optimization, LangGraph agent systems. +- Scope: Principal-level AI engineering work, including setting technical direction and solving the hardest AI/product problems across video learning, mobile constraints, AI output quality, and production cost. + +## Major Workstreams + +### AI Auto Notes + +- Problem: Learners needed automatic notes from video lectures. +- What Hamza built: LLM/Whisper-based automatic note generation from video. +- Technologies: OpenAI API, Whisper, Python/FastAPI, AWS, LangGraph. +- Result: Existing resume claims 12% video engagement lift and 25% overall retention boost. +- Evidence: Existing resume/CV and project pages. Metrics should be confirmed before heavy external use. + +### Mobile AI Innovation + +- Problem: Mobile platforms have constraints around video capture and processing. +- What Hamza built: React Native iOS/Android app features backed by server-side video processing using AWS Lambda and FFmpeg. +- Technologies: React Native, AWS Lambda, EventBridge, S3, FFmpeg, Python/FastAPI. +- Result: Existing resume claims 40% mobile retention increase. +- Evidence: Existing resume/CV. + +### AI Chat And RAG + +- Problem: Users needed to search and converse with their learning notes. +- What Hamza built: Conversational AI for notes using LangChain, Pinecone vector DB, semantic search, and streaming responses. +- Technologies: LangChain, Pinecone, embeddings, streaming UI. +- Result: Existing resume claims 25% session duration increase. +- Evidence: Existing resume/CV. + +### Real-Time Transcription + +- Problem: Users needed live transcription while learning from video. +- What Hamza built: WebRTC to Socket.io to speech-to-text pipeline. +- Technologies: WebRTC, Socket.io, Google Cloud Speech STT, Whisper/Groq provider evolution. +- Result: Existing resume claims 51% premium subscription increase and 90% transcription cost reduction. +- Evidence: Existing resume/CV and Stack Overflow impact mention. + +### Cross-Platform Authentication + +- Problem: Slid needed unified identity across web, Chrome Extension, React Native mobile, and Electron desktop. +- What Hamza built: Cross-platform authentication using AWS Cognito and custom storage/session patterns. +- Technologies: AWS Cognito, cookies/tokens, browser extension context, React Native, Electron. +- Result: Unified access and faster platform expansion. +- Evidence: Existing resume/CV and project pages. + +### Payment And Pricing + +- Problem: Slid needed international and Korean payment support with premium feature access. +- What Hamza built: Stripe and Payple multi-gateway integration with subscription and privilege control. +- Technologies: Stripe, Payple, backend payment APIs, webhooks. +- Result: Existing project pages claim 35% checkout improvement and 98%+ payment success. +- Evidence: `_projects/11_project.md`. + +### Export Systems + +- Problem: Users needed notes in multiple document/workspace formats. +- What Hamza built: Multi-format export to Word, PDF, Markdown, and Notion. +- Technologies: Document generation, Markdown/AST processing, Notion API. +- Result: Existing resume claims 96% accuracy and 75% support burden reduction. +- Evidence: Existing resume/CV. + +## Technical Stack + +- Frontend/mobile: React, React Native, Chrome Extension, Electron. +- Backend: Python, FastAPI, Node.js. +- AI: OpenAI API, Whisper, LangChain, LangGraph, Pinecone, embeddings. +- Realtime/media: WebRTC, Socket.io, FFmpeg. +- Cloud: AWS Lambda, EventBridge, S3, Cognito. +- Payments: Stripe, Payple, Apple In-App Purchase needs confirmation. + +## Resume Bullet Bank + +- Led AI transformation for Slid's 50K+ MAU EdTech platform, shipping LLM-powered auto-notes, AI chat, semantic search, and transcript correction systems. +- Built LLM/Whisper auto-notes from video lectures and helped drive measurable engagement and retention gains across learning workflows. +- Spearheaded React Native mobile AI features backed by server-side AWS Lambda and FFmpeg video processing to overcome mobile platform constraints. +- Designed LangGraph multi-agent transcript correction workflows with custom reasoning tools to improve AI-generated learning content quality. +- Built early RAG and conversational AI experiences for learning notes using LangChain, Pinecone, embeddings, and streaming responses. +- Pioneered real-time transcription pipeline using WebRTC, Socket.io, and cloud speech-to-text providers, then optimized provider strategy for major cost reductions. +- Built cross-platform authentication and payment systems across web, extension, mobile, and desktop, including AWS Cognito, Stripe, and Payple. + +## Site/Portfolio Angles + +- Strong anchor role: progression from early engineer to AI team lead. +- Good for showing repeated ownership across product, infrastructure, AI, mobile, and business metrics. +- Split into 2-4 portfolio projects if the site needs depth: Auto Notes, Mobile AI, RAG Chat, Payments/Auth. + +## Proof Links And Evidence + +- `_data/cv.yml` +- `_pages/about.md` +- `2025-10-25-hamza-kyamanywa-openai-resume.md` +- `_projects/10_project.md` through `_projects/15_project.md` contain detailed Slid-related project pages. + +## Needs Confirmation + +- Exact public-safe metrics and whether all can be published. +- Whether to use `Lead AI Engineer / Team Lead` as the public title on homepage/resume/LinkedIn while keeping `AI Engineer and Team Lead (Level 4)` as the internal/official progression. +- Which project pages are too inflated, outdated, or need pruning. diff --git a/career_evidence/roles/gogymi-textutor.md b/career_evidence/roles/gogymi-textutor.md new file mode 100644 index 0000000..1e89b4b --- /dev/null +++ b/career_evidence/roles/gogymi-textutor.md @@ -0,0 +1,134 @@ +# GoGymi / TexTutor + +## Metadata + +- Evidence status: Verified from contribution reports, final public wording needs review +- Public name: TexTutor +- Organization/client: GoGymi +- Role/title: Senior Full Stack Engineer / freelance technical lead +- Period: Jan 2025 - Sep 2025 +- Location: Remote, Switzerland +- Source repos: Needs confirmation +- Existing public artifacts: `_pages/about.md`, `_data/cv.yml`, `project_contributions/` + +## One-Line Positioning + +Delivered a full-stack AI-powered EdTech SaaS platform for Swiss teachers and students, spanning architecture, authentication, billing, multilingual UX, collaboration, and AI grammar correction. + +## Context + +TexTutor is described in existing site/resume materials as an AI-powered educational platform for Swiss teachers and students. Contribution reports describe Hamza as the primary/full-stack contributor across frontend, backend, billing, DevOps, authentication, and AI integrations. + +## Ownership + +- Full-stack platform architecture. +- Next.js/React frontend for teacher and student dashboards. +- Database schema and migration management. +- Authentication, email verification, and password reset flows. +- Stripe subscriptions, trials, billing dashboard, and webhook flows. +- Multilingual user interface and billing emails. +- AI grammar correction and feedback workflows. +- Rich text/collaboration workflows using Lexical and LiveBlocks. +- Deployment and local development infrastructure. + +## Major Workstreams + +### Architecture And Infrastructure + +- Problem: TexTutor needed a complete product foundation. +- What Hamza built: Next.js 15, React 19, PostgreSQL, Drizzle ORM, Supabase, TypeScript platform architecture. +- Technologies: Next.js, React, TypeScript, PostgreSQL, Drizzle ORM, Supabase, Docker. +- Result: Production/pre-launch platform foundation. +- Evidence: `project_contributions/TEXTUTOR_HAMZA_CONTRIBUTIONS.md`, `project_contributions/HAMZA_CONTRIBUTIONS2.md`. + +### Authentication And Security + +- Problem: Teachers/students needed secure onboarding and account management. +- What Hamza built: Authentication system with email verification, password reset, protected routes, sessions, and role-aware access. +- Technologies: JWT/session handling, bcrypt, email verification flows. +- Result: Complete user-management foundation. +- Evidence: Contribution reports. + +### Billing And Subscription System + +- Problem: GoGymi needed monetization with subscriptions, trials, usage tracking, and localized billing communication. +- What Hamza built: Stripe webhook integration, subscription lifecycle management, free trial logic, usage/credit tracking, billing UI, and multilingual billing emails. +- Technologies: Stripe, webhooks, Resend/email templates, TypeScript. +- Result: Complete billing and subscription system. +- Evidence: Contribution reports mention 15+ commits and specific billing work. + +### AI Grammar Correction + +- Problem: The product needed AI-powered language correction and educational feedback. +- What Hamza built: Grammar correction workflows with formatting preservation, rule/preset management, multilingual support, credit validation, and feedback UI. +- Technologies: AI providers need confirmation, Python integration, Next.js frontend. +- Result: Core AI product workflow for educational writing support. +- Evidence: Contribution reports. + +### Rich Text And Collaboration + +- Problem: Teachers/students needed assignment creation, feedback, commenting, and collaborative editing. +- What Hamza built: Lexical editor integration, LiveBlocks collaboration, comments/threading, share flows, notifications, and PDF export workflows. +- Technologies: Lexical, LiveBlocks, React, PDF export. +- Result: Collaborative assignment and feedback experience. +- Evidence: Contribution reports. + +### Internationalization + +- Problem: Swiss/European education workflows needed multilingual support. +- What Hamza built: i18n support across English, German, French, Spanish, and Italian, including billing emails. +- Technologies: i18n message files, localized email templates. +- Result: Multi-language product and communication layer. +- Evidence: Contribution reports and existing resume. + +## Metrics And Outcomes + +- 554-555 commits: + - Status: Verified from contribution reports. + - Evidence: `project_contributions/TEXTUTOR_HAMZA_CONTRIBUTIONS.md`, `project_contributions/HAMZA_CONTRIBUTIONS2.md`. + - Public wording: "Delivered 550+ commits across frontend, backend, billing, authentication, AI workflows, and deployment." +- 5 languages: + - Status: Verified from contribution reports/resume. + - Public wording: "Built multilingual product and billing workflows across EN/DE/ES/FR/IT." +- 43,000+ lines added: + - Status: Strong draft from contribution report. + - Public wording: Use only if Hamza wants commit-volume emphasis. + +## Technical Stack + +- Frontend: Next.js 15, React 19, TypeScript, Tailwind CSS, Radix UI, Lexical. +- Backend/data: PostgreSQL, Drizzle ORM, Supabase, server actions/API routes. +- Billing: Stripe, webhooks, subscriptions, free trials. +- Collaboration: LiveBlocks, comments, notifications. +- Email/i18n: Resend, multilingual templates, EN/DE/ES/FR/IT. +- AI: Grammar correction providers and Python integration need confirmation. +- DevOps: Docker, CI/CD needs confirmation. + +## Resume Bullet Bank + +- Delivered TexTutor, an AI-powered EdTech SaaS platform for Swiss teachers and students, across full-stack architecture, dashboards, authentication, billing, AI correction, and deployment workflows. +- Built complete Stripe subscription infrastructure with webhooks, trials, usage tracking, billing dashboards, and multilingual billing emails. +- Implemented multilingual product flows across English, German, French, Spanish, and Italian for user-facing UI and transactional communication. +- Integrated AI grammar correction workflows with formatting preservation, preset/rule management, usage validation, and educational feedback loops. +- Built rich assignment and feedback experiences with Lexical, LiveBlocks collaboration, comments, notifications, sharing, and PDF export. + +## Site/Portfolio Angles + +- Strong freelance/client delivery story. +- Good proof of end-to-end SaaS execution outside Slid/HoverNotes. +- Strong for roles needing product engineering, billing, i18n, AI features, and client ownership. + +## Proof Links And Evidence + +- `project_contributions/TEXTUTOR_HAMZA_CONTRIBUTIONS.md` +- `project_contributions/HAMZA_CONTRIBUTIONS2.md` +- `_data/cv.yml` +- `_pages/about.md` + +## Needs Confirmation + +- Whether public name should be `TexTutor`, `TextUtor`, or another spelling. +- Exact launch status. +- Source repo path and whether it can be referenced. +- Which AI providers were used. +- Public-safe wording for commit count and line count. diff --git a/career_evidence/roles/hovernotes.md b/career_evidence/roles/hovernotes.md new file mode 100644 index 0000000..859ce50 --- /dev/null +++ b/career_evidence/roles/hovernotes.md @@ -0,0 +1,177 @@ +# HoverNotes + +## Metadata + +- Evidence status: Repo-backed working draft, user-confirmed revenue/install metrics as of 2026-06-05 +- Public name: HoverNotes +- Organization/client: HoverNotes +- Role/title: Technical Founder and CEO +- Period: Jul 2025 - Present +- Location: Remote, United States / Seoul, South Korea needs confirmation +- Source repos: + - `private-local-workspace/hovernotes/hover-notes-web-app` + - Remote: `https://github.com/HoverNotes/hover-notes-web-app.git` + - Related repos inferred from docs, paths need confirmation: `hover-notes-extension`, `hover-notes-llm-api` +- Existing public artifacts: `_pages/about.md`, `_data/cv.yml`, `2025-10-25-hamza-kyamanywa-openai-resume.md` + +## One-Line Positioning + +Founded and built HoverNotes, a privacy-first AI video note-taking product with around 20K Chrome/Edge installs and over $10K/month revenue as of June 5, 2026, turning online videos into structured notes through a browser extension, a Next.js web app, local-first Obsidian workflows, and multimodal AI note generation. + +## Context + +HoverNotes is a privacy-first video learning platform. The product spans a Chrome/Edge browser extension, a Next.js web app, local file/vault integrations, payment and subscription systems, analytics, SEO/content infrastructure, and an AI backend integration for transcript, screenshot, and multimodal note generation. + +The local web app repo is currently a production Next.js app (`hover-notes`, version `1.17.17` in `package.json`) with active changelog entries through June 2026. + +## Ownership + +- Product strategy, architecture, implementation, and go-to-market. +- Chrome Extension for video detection/integration. +- Next.js web app for user and product workflows. +- Python FastAPI AI backend for note generation. +- AI pipelines using OpenAI GPT-4/Whisper, LangChain/LangGraph, and multilingual generation. +- Privacy-first local workflows and Obsidian integration. +- Pricing, payments, billing reliability, and region-specific monetization. +- SEO, content architecture, i18n/localization, and growth analytics. +- Production operations, debugging, telemetry, release/change management, and customer-facing stability work. + +## Major Workstreams + +### Product And Business + +- Problem: Learners need usable notes from video learning without losing ownership of their personal knowledge base. +- What Hamza built: Privacy-first AI video note-taking SaaS from concept to production product, including the web app, extension-facing workflows, pricing, payments, analytics, and growth loops. +- Technologies: Next.js, React, TypeScript, Drizzle/PostgreSQL, Stripe, Razorpay, PostHog, Vercel Analytics, Vercel Workflows. +- Result: User confirmed on 2026-06-05 that HoverNotes is over $10K/month revenue, not pure MRR because annual subscriptions are included, with around 20K installs across Chrome and Edge. Internal product analysis shows a 20-minute free tier with about 40% conversion among users who hit the paywall, 1,629 engaged AI Notes users, and 141,548 AI Notes minutes in one 30-day analysis window. +- Evidence: User statement on 2026-06-05 for revenue/install metrics; `docs/product-research/2026-04-06-ai-notes-funnel-analysis.md` for funnel/product analytics. Confirm exact wording and precision before publishing beyond this draft. + +### AI Note Generation + +- Problem: Video content needs to become structured, useful study notes across languages. +- What Hamza built: Extension/web-app AI Notes flow that records learning sessions, streams status into a Lexical editor, gates legacy versus modern backend pipelines by extension version, and supports transcript-only, screenshot-aware, and multimodal note generation paths. +- Technologies: Lexical, Zustand, TypeScript, PostHog telemetry, backend AI APIs, Whisper/Groq transcription needs confirmation, OpenAI/GPT-4 needs confirmation, LangGraph/LangChain needs confirmation. +- Result: Production AI Notes workflow with active state UI, language selection, Mermaid preference, capture recovery, warning/error telemetry, async v10_1 job polling, and AI-generated editor blocks. +- Evidence: `CHANGELOG.md` entries for PRs #541-#589 especially AI Notes v10_1 async jobs, audio-only mode, Mermaid preference, capture recovery, diagnostics, telemetry, and UI improvements; `components/lexical-editor/plugins/AINotesPlugin/index.tsx`; `components/lexical-editor/nodes/TranscriptionNode.tsx`; `lib/ai-notes-request.ts`; `lib/ai-notes-pipeline.ts`. + +### Multimodal Screenshots Pipeline + +- Problem: Backend-side video frame extraction is slower, more expensive, and less visually faithful than using captured frames from the browser. +- What Hamza built: Cross-repo v10 AI Notes plan where the extension captures scene-change screenshots, the web app deduplicates/ranks/downsized screenshots, and the backend receives transcript plus selected segment images for multimodal generation and critique. +- Technologies: Browser capture, image deduplication, dHash, MediaPipe similarity needs confirmation, screenshot quality scoring, backend v10 payload contracts. +- Result: Flag-gated pipeline intended to reduce latency/cost and improve visual grounding by sending only high-value screenshots. +- Evidence: `docs/projects/ai-notes-audio-screenshots.md`; `components/lexical-editor/utils/imageDedup.ts`; `components/lexical-editor/utils/aiCaptureImageResolver.ts`; `lib/ai-notes-request.ts`; tests in `lib/__tests__/ai-notes-request.test.ts` and `components/lexical-editor/utils/aiCaptureImageResolver.test.ts`. + +### Universal Video Compatibility + +- Problem: Learners use many video platforms, not only YouTube. +- What Hamza built: Extension-facing editor and product flows designed for videos across YouTube, Udemy, Coursera, Bilibili, and other learning platforms; app copy and analytics explicitly track platform-specific behavior and drop-off. +- Technologies: Chrome/Edge extension APIs needs confirmation, iframe postMessage API, browser media workflows, TypeScript, React. +- Result: Broader platform compatibility and targeted fixes for platform-specific UX such as YouTube ad pause/resume handling, ad muting/speed restoration, and extension version nudges. +- Evidence: `HOVERNOTES_ARCHITECTURE.md`; `CHANGELOG.md` YouTube ad pause/silence/speed entries; `docs/product-research/2026-04-06-ai-notes-funnel-analysis.md` domain analysis; `lib/extension-store-versions.ts`; `lib/extension-store-urls.ts`. + +### Privacy And Obsidian Integration + +- Problem: Users want notes in their own local knowledge systems. +- What Hamza built: Local-first editor flow with direct Obsidian vault saving, file-management UX, local screenshot insertion, note history, auto-save/destructive-save guardrails, and vault reconnect/recovery logic. +- Technologies: File System Access API needs confirmation, iframe postMessage bridge, Lexical editor, Markdown export, local screenshot handling. +- Result: User-owned notes and images stored in local vault workflows rather than permanent cloud storage by default. +- Evidence: `HOVERNOTES_ARCHITECTURE.md`; `docs/FILE-MANAGEMENT-ARCHITECTURE.md`; `CHANGELOG.md` entries for note history, save recovery, destructive save guard, screenshot capture, and vault feedback tracking. + +### Monetization And Billing Infrastructure + +- Problem: A global SaaS needs payments that match regional buyer behavior and can handle recurring access reliably. +- What Hamza built: Multi-region pricing and payment system covering Stripe for global card payments, WeChat Pay/Alipay one-time passes for China, Razorpay/UPI for India, subscription status APIs, webhook handling, and Vercel Workflow-based recurring billing where Razorpay lacks native cross-border subscriptions. +- Technologies: Stripe, Razorpay, Vercel Workflows, Drizzle/PostgreSQL, R2/invoice storage needs confirmation. +- Result: Production pricing tiers for Global, East Asia, India, and China; region-specific payment options; retry/cancel flows; subscription access cleanup on terminal Stripe statuses; Razorpay workflow health/debug docs. +- Evidence: `docs/PRICING-TIERS.md`; `app/pricing/pricing-data.ts`; `lib/geo/countries-pricing-data.ts`; `app/api/checkout/route.ts`; `app/api/stripe/webhook/route.ts`; `app/api/razorpay/*`; `app/workflows/razorpay-recurring-billing.ts`; `docs/razorpay-recurring-billing-workflow.md`; `CHANGELOG.md` PRs #565-#566 and #593. + +### Internationalization And Market Expansion + +- Problem: Video learners are global, and the product needs localized UI, pricing, SEO, and content. +- What Hamza built: Locale system, translated UI copy across 10 supported locales, blog translation workflow, language selectors for public and extension experiences, region-specific Chinese homepage/pricing copy, and AI Notes language shortcuts. +- Technologies: Next.js App Router i18n routes, JSON locale files, translation import/export scripts, database-backed blog translations needs confirmation. +- Result: Supported locales include Korean, Chinese, Japanese, Italian, Portuguese, Russian, German, Spanish, Vietnamese, and French; public and extension copy kept aligned across locales. +- Evidence: `lib/i18n/locales/*.json`; `docs/TRANSLATION-WORKFLOW.md`; `app/[lang]/`; `CHANGELOG.md` PRs #583-#586 and #584 Chinese marketing copy. + +### SEO, Content, And Discoverability + +- Problem: A new product needs search visibility, canonical hygiene, structured data, localized content, and crawl control across production/staging. +- What Hamza built: Dynamic metadata/canonical helpers, sitemap/robots handling, schema markup for blog collection/articles/breadcrumbs/videos, multilingual hreflang-style metadata needs confirmation, and SEO documentation/audits. +- Technologies: Next.js metadata API, JSON-LD, MDX/blog system, robots/sitemap routes. +- Result: Canonical domain consistency for `hovernotes.io`, crawler controls for staging, schema-rich blog content, localized marketing/blog pages. +- Evidence: `lib/seo/metadata-helpers.ts`; `app/robots.ts`; `app/sitemap.ts`; `docs/SEO-SCHEMA-MARKUP.md`; `SEO-AUDIT-CURRENT-STATE.md`; `CHANGELOG.md` Chinese SEO copy and language selector entries. + +### Production Reliability And Observability + +- Problem: AI video capture is failure-prone across browsers, devices, store versions, ad states, network paths, and backend pipelines. +- What Hamza built: Telemetry and recovery systems for AI Notes sessions, extension diagnostics, media payload sizes, warning correlation, active-session heartbeat, fallback upload paths, old-extension routing, and customer-facing error states. +- Technologies: PostHog, browser/extension diagnostics, structured event payloads, tests, Vercel logs/workflows. +- Result: Better ability to detect active sessions, debug weak machines, correlate warning chunks to sessions, route older extensions to compatible backends, and avoid stuck editor states. +- Evidence: `CHANGELOG.md` PRs #556-#579 and #593; `docs/vercel-workflows-guide.md`; `lib/__tests__/ai-notes-request.test.ts`; `lib/__tests__/extension-store-versions.test.ts`. + +## Technical Stack + +- Frontend: Next.js 15, React 19, TypeScript, Tailwind CSS, Radix UI, Lexical editor, Zustand. +- Browser/extension: Chrome/Edge extension integration, iframe postMessage bridge, video capture workflows. Exact extension repo/path needs confirmation. +- Backend/API: Next.js route handlers, Drizzle ORM, PostgreSQL/Supabase, NextAuth.js, Python FastAPI AI backend needs confirmation. +- AI: OpenAI/GPT-4, Whisper/Groq transcription, LangChain, LangGraph, multimodal screenshot pipelines. Backend provider details need confirmation before final public wording. +- Payments/monetization: Stripe, Razorpay, Vercel Workflows, regional pricing, one-time passes, subscription webhooks. +- Analytics/ops: PostHog, Vercel Analytics, Vercel Speed Insights, internal product analysis scripts/docs. +- Growth/content: MDX/blog system, JSON-LD schema, sitemap/robots, locale JSON files, translation scripts. +- Integrations: Obsidian vault workflows, YouTube/Udemy/Coursera/Bilibili and other video platforms. + +## Resume Bullet Bank + +- Founded HoverNotes, a privacy-first AI video note-taking SaaS, growing it to over $10K/month revenue and around 20K Chrome/Edge installs while owning product strategy, architecture, implementation, and go-to-market. +- Built the production Next.js/React web app for the extension editor, subscriptions, pricing, localization, SEO, analytics, and local-first note-taking workflows. +- Architected AI Notes flows that coordinate browser capture, editor state, backend routing, asynchronous job polling, telemetry, and recovery across legacy and modern AI pipelines. +- Designed a multimodal video-in, notes-out pipeline that captures and deduplicates key frames client-side before sending selected visual context to the backend for AI note generation. +- Built local-first Obsidian vault workflows, including file save/reconnect paths, screenshot handling, note history, and guardrails against destructive auto-save behavior. +- Implemented region-aware monetization across Stripe for global card payments, WeChat Pay/Alipay one-time passes for China, and Razorpay/UPI for India, including Vercel Workflow-based recurring billing and webhook/subscription access controls. +- Used product analytics to diagnose AI Notes activation and paywall behavior, including a 30-day analysis showing about 40% conversion among users who reached the 20-minute paywall. Confirm before public use. +- Built multilingual product and content infrastructure across 10 supported locales, including localized UI copy, translated blog workflows, and market-specific Chinese SEO/pricing copy. +- Improved search visibility with canonical metadata, sitemap/robots handling, schema.org JSON-LD for blog/articles/breadcrumbs/videos, and localized marketing content. +- Added production observability for AI Notes sessions, media payloads, extension diagnostics, warning correlation, and active-session heartbeat tracking to debug cross-browser/device failures. +- Maintained release discipline through internal changelog/versioning, focused PRs, tests around AI Notes contracts and extension-version gates, and production incident docs. + +## Site/Portfolio Angles + +- Founder story: technical depth plus product and customer ownership. +- AI product case study: local-first AI, browser extension, video processing, and knowledge management. +- Strong for roles needing 0-to-1 ownership, AI systems architecture, or founder empathy. +- Monetization case study: region-aware pricing, Stripe/Razorpay/China payment methods, billing reliability, and paywall analytics. +- Growth case study: SEO/i18n/blog infrastructure plus measurable funnel analysis. +- Reliability case study: turning messy browser video capture into observable, recoverable product infrastructure. + +## Proof Links And Evidence + +- `_pages/about.md` +- `_data/cv.yml` +- `2025-10-25-hamza-kyamanywa-openai-resume.md` +- `private-local-workspace/hovernotes/hover-notes-web-app/HOVERNOTES_ARCHITECTURE.md` +- `private-local-workspace/hovernotes/hover-notes-web-app/CHANGELOG.md` +- `private-local-workspace/hovernotes/hover-notes-web-app/package.json` +- `private-local-workspace/hovernotes/hover-notes-web-app/docs/projects/ai-notes-audio-screenshots.md` +- `private-local-workspace/hovernotes/hover-notes-web-app/docs/product-research/2026-04-06-ai-notes-funnel-analysis.md` +- `private-local-workspace/hovernotes/hover-notes-web-app/docs/PRICING-TIERS.md` +- `private-local-workspace/hovernotes/hover-notes-web-app/docs/razorpay-recurring-billing-workflow.md` +- `private-local-workspace/hovernotes/hover-notes-web-app/docs/SEO-SCHEMA-MARKUP.md` +- `private-local-workspace/hovernotes/hover-notes-web-app/docs/TRANSLATION-WORKFLOW.md` +- `private-local-workspace/hovernotes/hover-notes-web-app/lib/seo/metadata-helpers.ts` +- `private-local-workspace/hovernotes/hover-notes-web-app/lib/i18n/locales/` +- `private-local-workspace/hovernotes/hover-notes-web-app/app/workflows/razorpay-recurring-billing.ts` +- `private-local-workspace/hovernotes/hover-notes-web-app/components/lexical-editor/plugins/AINotesPlugin/index.tsx` +- `private-local-workspace/hovernotes/hover-notes-web-app/components/lexical-editor/nodes/TranscriptionNode.tsx` + +## Needs Confirmation + +- Current status of HoverNotes. +- Exact revenue wording and whether to publish publicly: over $10K/month revenue, not MRR because annual subscriptions are included. +- Exact Chrome/Edge install count and whether to round publicly as 20K+ installs. +- Exact job title to use publicly: `Technical Founder and CEO`, `Founder`, `Founder/Engineer`, or another framing. +- Location phrasing. +- Extension repo path, backend repo path, and whether they can be named. +- Backend provider details: OpenAI/GPT-4, Whisper/Groq, LangChain, LangGraph, FastAPI. +- Whether internal funnel metrics can be used publicly, and at what precision. +- Whether period should still be `Present`. diff --git a/career_evidence/roles/jonjabird.md b/career_evidence/roles/jonjabird.md new file mode 100644 index 0000000..55e19a9 --- /dev/null +++ b/career_evidence/roles/jonjabird.md @@ -0,0 +1,167 @@ +# JonjaBird + +## Metadata + +- Evidence status: Verified local business workspace; public framing and metrics need confirmation +- Public name: JonjaBird +- Organization/client: JonjaBird / HOVERNOTES LLC product line needs confirmation +- Role/title: Founder, product operator, and automation engineer +- Period: Mar 2026 - Present needs confirmation +- Location: Seoul, South Korea +- Source workspace: + - `private-local-workspace/inmate-letters-business` +- Existing public artifacts: Not yet integrated into site/resume +- Private/internal artifacts: + - `business-docs/` + - `client-accounts/` + - `templates/` + - `scripts/` + +## One-Line Positioning + +Founded and operationalized JonjaBird, a hybrid physical-digital mail and support service for foreign inmates in South Korea and their overseas sponsors, combining service design, pricing strategy, legal/privacy docs, client ledgers, and Python/WeasyPrint document automation. + +## Context + +JonjaBird serves a communication gap for non-Korean inmates in Korean correctional facilities. Inmates have no internet access and international families often cannot reliably manage Korean prison mail, printing, address formatting, translation, payments, or account visibility. + +The current operating model is a manual MVP: inbound inmate letters are received and scanned, sponsors receive digital copies, sponsor replies are printed and mailed back through Korean post, and related charges/deposits are tracked through client account ledgers and generated PDFs. The workspace also contains the spec for a future sponsor/admin portal that would migrate the Excel ledger and PDF workflow into a Next.js + FastAPI system. + +## Ownership + +- Founder/operator for the service concept, launch plan, customer workflows, and operating docs. +- Business analysis across market sizing, competitor comparison, pricing, unit economics, legal/privacy posture, and service catalog design. +- Customer-facing materials: intro letters, service forms, terms, tone guide, multilingual one-pagers, receipts, deposit confirmations, account statements, and typed/translated client documents. +- Operations system for client account ledgers, receipt/deposit/statement numbering, credit balances, service charges, and auditability. +- Python PDF automation using WeasyPrint, openpyxl, shared font handling, and reusable script conventions. +- Future platform architecture/spec for sponsor portal, admin queue, deposits, ledgers, receipts/statements, payments, storage, and document-processing service boundaries. + +## Major Workstreams + +### Business And Service Design + +- Problem: Foreign inmates in Korea need a reliable bridge to sponsors abroad, and sponsors need proof that letters, services, and funds are being handled transparently. +- What Hamza built: A full operating model for letter forwarding, AI research, typing/transcription, court-document translation, photo printing, money transfer support, credits, statements, and heavy printed-package pricing. +- Technologies/process: Business docs, pricing model, service catalog, changelog decision log, operational playbooks, customer-facing forms, and private client account trackers. +- Result: A real service system with clear workflows, pricing logic, grandfathering rules, package rules, and customer-facing documentation. +- Evidence: `business-docs/01-business-overview.md`, `03-pricing-model.md`, `05-operational-plan.md`, `10-onboarding-playbook.md`, `12-service-catalog.md`, `CHANGELOG.md`. + +### Pricing, Unit Economics, And Trust Model + +- Problem: Sponsors need transparent, itemized spending, while the business needs pricing that covers Korean postage, printing, handling, and operational risk. +- What Hamza built: Pricing revisions that moved the service away from commodity postage, bundled tracked express mail into core services, introduced a credit/balance model, separated money transfers from JonjaBird credits, and added a large printed-package rule for high-page-count packets. +- Technologies/process: Pricing model spreadsheet, markdown pricing model, change-log governance, customer-specific grandfathering, ledger/statement alignment. +- Result: Current pricing model documents per-service margins, market comparisons, order examples, overage rules, and the distinction between service credits and inmate-bank money transfers. +- Evidence: `business-docs/03-pricing-model.md`, `business-docs/CHANGELOG.md`, `business-docs/pricing-revenue-model.xlsx`. + +### Document Automation And Audit Trail + +- Problem: A sensitive mail/payment service needs clean proof of charges, deposits, balances, and account history without hand-making every PDF. +- What Hamza built: Python + WeasyPrint PDF templates for service receipts and credit deposit confirmations, plus a statement generator that reads `.xlsx` credit trackers and produces branded account statements. +- Technologies: Python, WeasyPrint, openpyxl, HTML/CSS PDF layouts, local font stack, file-path-safe script conventions. +- Result: Reusable generation workflow for service receipts (`JB-2026-XXXX`), deposit confirmations (`DEP-2026-XXXX`), and statements (`STMT-2026-XXXX`) with output folders, numbering rules, and PDF gotchas documented. +- Evidence: `scripts/README.md`, `templates/generate_receipt.py`, `templates/generate_deposit_confirmation.py`, `scripts/build_statement.py`, `client-accounts/receipts/`, `client-accounts/deposit-confirmations/`, `client-accounts/statements/`. + +### Client Account Operations + +- Problem: Sponsors and inmates need balances and service activity that reconcile with payments and PDF artifacts. +- What Hamza built: Client account trackers and a workflow that keeps deposits and charges as separate ledger rows, aligns balances with receipts/deposit confirmations, and produces statements from the same source data. +- Technologies/process: Excel `.xlsx` credit trackers, generated PDFs, sequential document numbering, manual deposit confirmation flow. +- Result: Private client account files exist for ABU Donzo, Giedrius Voronovic, and Jacek Polanski, with receipt outputs visible through June 2026. +- Evidence: `client-accounts/abu-donzo-credit-tracker.xlsx`, `client-accounts/giedrius-voronovic-credit-tracker.xlsx`, `client-accounts/jacek-polanski-credit-tracker.xlsx`, `client-accounts/receipts/`. + +### AI-Assisted Prisoner Services + +- Problem: Inmates lack internet access and often need research, typing, translation, or document prep that is difficult to request from prison. +- What Hamza built: Service definitions and one-off deliverables for AI research, handwritten-to-typed documents, legal/court document translation, address overlays, bilingual questionnaire translation, and prison-mail packages. +- Technologies: AI-assisted drafting/transcription/translation, Python PDF generation, bilingual formatting, Korean/English document workflows. +- Result: The workspace contains concrete generated client deliverables, including typed legal/court documents, Korean translations, research reports, and bilingual client materials. +- Evidence: `business-docs/12-service-catalog.md`, `client-accounts/abu-donzo/`, `client-accounts/typed-letters/`, `scripts/build_sam_research_js.py`, `scripts/build_sam_sexed_translation.py`, `scripts/generate_clemency_letter_abu_danzo.py`, `scripts/generate_clemency_letter_abu_danzo_korean.py`. + +### Future Sponsor/Admin Platform Spec + +- Problem: Manual operation works for seed clients but will not scale once sponsor submissions, deposits, scans, charges, and statements grow. +- What Hamza built: A living web app spec for sponsor signup, inmate profiles, letter/photo submission, pay-at-send, credit loading, manual deposit fallback, admin queue, scanned replies, ledgers, receipts/statements, and future OCR/translation/AI pipelines. +- Technologies planned: Next.js App Router, TypeScript, FastAPI, WeasyPrint, Postgres, Prisma, S3-compatible storage, Stripe, Wise/manual deposit fallback, Resend/Postmark, Vercel, Fly.io/Railway. +- Result: The spec preserves compatibility between today's `.xlsx` ledger/PDF scripts and the eventual DB-backed portal. +- Evidence: `business-docs/19-web-app-spec.md`, especially sections 3, 4, 5, 6, and 8. + +## Metrics And Outcomes + +- Generated service receipts: + - Status: Verified local artifacts and script README numbering. + - Evidence: `client-accounts/receipts/`, `scripts/README.md`. + - Public wording: "Built an auditable receipt workflow with sequential service receipts through `JB-2026-0016`." Confirm whether to expose numbering publicly. +- Deposit confirmations: + - Status: Verified workflow; artifact visibility varies by local folder state. + - Evidence: `templates/generate_deposit_confirmation.py`, `scripts/README.md`, `client-accounts/deposit-confirmations/`. + - Public wording: "Separated service receipts from credit deposit confirmations to make sponsor balances easier to audit." +- Client account system: + - Status: Verified local files. + - Evidence: `client-accounts/*-credit-tracker.xlsx`. + - Public wording: "Maintained client credit ledgers that reconcile deposits, charges, balances, receipts, and statements." +- Pricing/unit economics: + - Status: Verified docs; exact public metrics need approval. + - Evidence: `business-docs/03-pricing-model.md`, `CHANGELOG.md`. + - Public wording: "Designed pricing and unit economics for a specialized mail-and-document service with tracked Korean postal delivery." +- Customer/client count: + - Status: Needs confirmation before public use. + - Evidence: `client-accounts/` suggests multiple named clients, but names are private. + - Public wording: "Supported early paying clients" only if Hamza confirms wording and privacy posture. + +## Technical Stack + +- Business/docs: Markdown, Excel `.xlsx`, generated PDFs, changelog-driven decision log. +- Automation: Python, WeasyPrint, openpyxl, HTML/CSS for PDF layouts. +- Document generation: A5 receipts/deposit confirmations, account statements, service forms, one-pagers, bilingual letters. +- AI/document work: AI-assisted research, transcription, translation, legal/court document prep. +- Payments/ledger: Manual deposit confirmation, Wise, Stripe planned/available via existing infrastructure, credit balances, money-transfer fee model. +- Future app: Next.js App Router, TypeScript, FastAPI, Postgres, Prisma, S3-compatible storage, Stripe Checkout, Resend/Postmark, Vercel, Fly.io/Railway. +- Operations: Korean post workflows, prison mail addressing, package restrictions, sponsor/inmate service agreements, privacy/data-retention rules. + +## Resume Bullet Bank + +- Founded JonjaBird, a hybrid physical-digital mail and support service helping foreign inmates in South Korea communicate with sponsors abroad through scanned letters, printed replies, payments, and document support. +- Designed the end-to-end operating system for inmate intake, sponsor onboarding, letter forwarding, service requests, credit balances, money transfers, statements, and customer-facing terms. +- Built Python/WeasyPrint automation for branded service receipts, credit deposit confirmations, and account statements, with sequential numbering and reusable output conventions. +- Created client credit-ledger workflows that separate deposits from service charges and reconcile balances against receipts, deposit confirmations, and generated statements. +- Developed pricing and unit-economics models for letter forwarding, AI research, transcription, translation, photo printing, money transfers, and large printed mail packages. +- Authored a future sponsor/admin portal spec mapping today's Excel/PDF workflow into a Next.js + FastAPI architecture with Stripe payments, admin queueing, Postgres ledgers, file storage, and WeasyPrint PDF endpoints. +- Produced AI-assisted research, transcription, translation, and bilingual document deliverables for real inmate support workflows under strict accuracy and privacy constraints. +- Established operational governance through a changelog, service catalog, tone guide, legal/privacy terms, onboarding playbook, and PDF generation README. + +## Site/Portfolio Angles + +- Founder/operator case study: turning a painful, real-world communication gap into a working service with docs, money flows, and artifacts. +- Product-systems case study: operational workflows first, then a future platform spec that cleanly maps manual state into software. +- Automation case study: using Python PDF generation and spreadsheets to make a sensitive manual service auditable before building a full app. +- AI-for-operations case study: practical AI research/transcription/translation for people with no internet access, framed carefully because some content is legal-adjacent. +- Strong for roles needing founder empathy, complex workflow design, compliance-aware product thinking, and boring-but-critical internal tools. + +## Proof Links And Evidence + +- `private-local-workspace/inmate-letters-business/business-docs/01-business-overview.md` +- `private-local-workspace/inmate-letters-business/business-docs/03-pricing-model.md` +- `private-local-workspace/inmate-letters-business/business-docs/05-operational-plan.md` +- `private-local-workspace/inmate-letters-business/business-docs/10-onboarding-playbook.md` +- `private-local-workspace/inmate-letters-business/business-docs/12-service-catalog.md` +- `private-local-workspace/inmate-letters-business/business-docs/16-terms-and-conditions.md` +- `private-local-workspace/inmate-letters-business/business-docs/17-style-and-tone-guide.md` +- `private-local-workspace/inmate-letters-business/business-docs/19-web-app-spec.md` +- `private-local-workspace/inmate-letters-business/business-docs/CHANGELOG.md` +- `private-local-workspace/inmate-letters-business/scripts/README.md` +- `private-local-workspace/inmate-letters-business/templates/generate_receipt.py` +- `private-local-workspace/inmate-letters-business/templates/generate_deposit_confirmation.py` +- `private-local-workspace/inmate-letters-business/scripts/build_statement.py` +- `private-local-workspace/inmate-letters-business/client-accounts/` + +## Needs Confirmation + +- Whether JonjaBird should be public on the site/resume at all, given privacy and legal-adjacent sensitivity. +- Public role title: `Founder`, `Founder/Operator`, `Product Founder`, `Founder and Automation Engineer`, or another framing. +- Period and whether to present it as active. +- Whether to name HOVERNOTES LLC as the operating entity. +- Which client count/revenue/payment metrics are safe to publish. +- Whether to mention "inmates," "incarcerated people," or a softer public phrase depending on audience. +- Whether to include client-facing legal/court-document work on a resume, or keep it as private evidence only. +- Whether to publish any pricing details or only describe pricing/unit-economics work generally. diff --git a/career_evidence/roles/quest-pdf-service.md b/career_evidence/roles/quest-pdf-service.md new file mode 100644 index 0000000..45da82a --- /dev/null +++ b/career_evidence/roles/quest-pdf-service.md @@ -0,0 +1,108 @@ +# Quest PDF Service + +## Metadata + +- Evidence status: Verified local repo and user-stated CTO/co-founder role, dates need confirmation +- Public name: Quest PDF Service +- Organization/client: Quest +- Role/title: CTO and Co-founder / PDF infrastructure owner +- Period: 2026 - Present needs confirmation +- Location: Needs confirmation +- Source repo: `private-local-workspace/quest/quest-pdf-service` +- Existing public artifacts: Not yet integrated into site/resume + +## One-Line Positioning + +Led the PDF infrastructure for Quest, building a production microservice that turns quote JSON into branded, paginated A4 safari quote PDFs, previews, and stored outputs in Cloudflare R2. + +## Context + +Quest needs professional quote documents for safari operators and travelers. The web app sends structured quote data to this service, which renders branded templates, paginates content, generates PDFs in headless Chrome, and uploads the result to Cloudflare R2. + +The PDF service is part of the broader Quest CTO/co-founder role and pairs with the web app's quote builder, page settings, preview/edit drawers, quote email workflow, R2 asset handling, and output-level monetization. + +## Ownership + +- Express/TypeScript service for PDF generation endpoints. +- Handlebars template architecture with modular page partials. +- Browser-based pre-pagination engine for A4 document layout. +- Puppeteer orchestration for PDF generation. +- Cloudflare R2 upload/delete storage integration. +- Per-page preview and screenshot-generation support. +- Signature background-removal endpoint for handwritten signatures. +- Docker/Hetzner deployment workflow. + +## Major Workstreams + +### Quote PDF Rendering + +- Problem: Quest needs pixel-perfect safari quote documents from structured app data. +- What Hamza built: JSON-to-HTML-to-PDF pipeline using Handlebars templates and Puppeteer. +- Technologies: Node.js 20, TypeScript, Express, Handlebars, Puppeteer. +- Result: Quote data can be converted into branded PDFs via `/generate-pdf` and supporting preview flows. +- Evidence: `quest-pdf-service/README.md`, `src/server.ts`, `src/pdf-generator.ts`, `template/`. + +### Pre-Pagination Engine + +- Problem: Dynamic itinerary and pricing content can overflow fixed A4 pages. +- What Hamza built: In-browser pagination that measures rendered content, splits overflowing pages, adds continuation headers, and controls footers. +- Technologies: Puppeteer, browser DOM measurement, TypeScript. +- Result: More reliable document output for variable-length safari quotes. +- Evidence: `quest-pdf-service/README.md`, `src/paginator.ts`. + +### Template System + +- Problem: Quote PDFs need reusable pages and configurable order. +- What Hamza built: Modular Handlebars partials for cover, summary, itinerary, pricing, terms, vehicles, about, and back cover, with theme support. +- Technologies: Handlebars, CSS variables, TypeScript helpers. +- Result: Reusable branded document system with per-company customization. +- Evidence: `quest-pdf-service/README.md`, `src/template-engine.ts`, `template/partials`. + +### Storage And Environments + +- Problem: Generated PDFs need durable public storage with separate dev/prod targets. +- What Hamza built: Cloudflare R2 upload/delete integration with company folder structure and request-level storage target switching. +- Technologies: Cloudflare R2, S3-compatible SDK, Express. +- Result: Generated PDFs are stored and returned as public URLs, with dev/prod bucket separation. +- Evidence: `quest-pdf-service/README.md`, `src/r2.ts`. + +### Signature Processing + +- Problem: Operators may need clean handwritten signatures in quote PDFs. +- What Hamza built: `/process-signature` endpoint for signature background removal. +- Technologies: Node.js image processing pipeline. Exact library needs confirmation. +- Result: Better-looking branded PDFs with clean signature assets. +- Evidence: `quest-pdf-service/README.md`, `src/signature.ts`. + +## Technical Stack + +- Backend: Node.js 20, TypeScript, Express. +- Rendering: Puppeteer, Handlebars, browser DOM measurement. +- Storage: Cloudflare R2. +- Deployment: Docker, Hetzner. +- Templates: HTML/CSS, Handlebars partials, CSS variables. + +## Resume Bullet Bank + +- Built Quest's PDF generation microservice, converting quote JSON into branded A4 safari quote PDFs with Puppeteer, Handlebars, and TypeScript. +- Designed a pre-pagination engine that measures rendered content in headless Chrome, splits overflowing itinerary/pricing sections, and injects continuation headers and controlled footers. +- Implemented Cloudflare R2 upload/delete workflows with dev/prod storage targeting and company-scoped folder organization. +- Created modular quote templates with configurable page order, company theming, full-bleed covers, itinerary pages, pricing pages, terms, vehicles, about pages, back covers, and per-page previews. + +## Site/Portfolio Angles + +- Strong case study for document-generation infrastructure and layout engineering. +- Good complement to Quest platform entry: web app plus specialized PDF microservice. +- Useful for showing backend craft beyond standard CRUD apps. + +## Proof Links And Evidence + +- Local repo: `private-local-workspace/quest/quest-pdf-service` +- Repo README documents features, endpoints, architecture, storage, and deployment. + +## Needs Confirmation + +- Exact start/end period. +- Public role title: user stated CTO and co-founder. +- Whether generated PDFs are live in production and at what volume. +- Signature image-processing implementation details and any measurable quality/support outcomes. diff --git a/career_evidence/roles/quest-platform.md b/career_evidence/roles/quest-platform.md new file mode 100644 index 0000000..cb00e19 --- /dev/null +++ b/career_evidence/roles/quest-platform.md @@ -0,0 +1,135 @@ +# Quest Safari Quotation Platform + +## Metadata + +- Evidence status: Verified local repo and user-stated CTO/co-founder role and 3-person team leadership, exact dates need confirmation +- Public name: Quest Safari Quotation Platform +- Organization/client: Quest +- Public URL: https://questheaven.com +- Role/title: CTO and Co-founder +- Period: 2026 - Present needs confirmation +- Location: Needs confirmation +- Source repo: `private-local-workspace/quest/quest-web-app` +- Repo remote: `https://github.com/questdev-canine/quest-web-app.git` +- Existing public artifacts: Not yet integrated into site/resume + +## One-Line Positioning + +Co-founded and leads technology for Quest (https://questheaven.com), a production Next.js platform that helps Uganda-based tour operators create accurate safari quotations, manage itinerary data, deliver branded quotes, and run operator-facing quote workflows. + +## Context + +Quest is a web application for Uganda-based tour operators. The platform supports Quest/admin operations, operator dashboards, company onboarding, document verification, content libraries, travel requests, itinerary building, quote generation, quote email delivery, PDF previews/exports, tourist tracking, billing, and analytics. + +## Ownership + +- Platform application development across Next.js App Router, TypeScript, Tailwind CSS, shadcn/ui, PostgreSQL/Supabase, Drizzle ORM, NextAuth v5, and Resend. +- Admin panel systems for managing platform data such as destinations, activities, hotels/lodges, transport, companies, users, settings, categories, and geography. +- Operator dashboard workflows for travel requests, itinerary creation, quote generation, quote email delivery, PDF preview/export, content libraries, analytics, billing, and tourist tracking. +- Product and architecture decisions around company-scoped content, trial/subscription gating, quote output monetization, Stripe billing, exchange rates, PDF infrastructure, and release workflow. +- Leadership of a 3-person engineering/product team. +- Deployment workflow and branch promotion process for shared QA, personal QA branches, and production releases. + +## Major Workstreams + +### Platform Architecture + +- Problem: Tour operators need fast, accurate, structured safari quotation workflows. +- What Hamza built: Next.js 16 App Router application with typed database access, role-based surfaces, company onboarding, operator dashboards, content libraries, quote workflows, analytics, and billing surfaces. +- Technologies: Next.js 16, React 19, TypeScript, Tailwind CSS 4, shadcn/ui, Drizzle ORM, PostgreSQL/Supabase. +- Result: Production platform foundation for Quest's quotation operations and SaaS launch path. +- Evidence: `quest-web-app/README.md`, `package.json`, `src/app`, `src/db`, `docs/progress/hamza/CURRENT.md`. + +### Authentication And Roles + +- Problem: Quest requires separate access patterns for internal admins, tour operators, and future partner users. +- What Hamza built: Role-aware app structure using NextAuth v5 and protected app/admin surfaces. +- Technologies: NextAuth v5, PostgreSQL, Drizzle ORM, middleware/server routes. +- Result: Admin and operator workflows can be separated cleanly. +- Evidence: `quest-web-app/README.md` role table and project structure. + +### Quote Operations + +- Problem: Operators need to convert travel requests into itineraries and quotes. +- What Hamza built: Operator dashboard architecture for requests, tourists, itineraries, pricing, quote generation, PDF preview/export, and quote email delivery. +- Technologies: Next.js App Router, server actions/API routes, database models. +- Result: Workflow support for quote creation and management. +- Evidence: `quest-web-app/README.md`; `src/app/app/quotes`; `src/app/app/requests`; `src/app/app/tourists`; `docs/progress/hamza/CURRENT.md`. + +### Product Strategy, Monetization, And Billing + +- Problem: Quest needs a launch model that gives operators enough freedom to build quotes while monetizing professional outputs. +- What Hamza built: Free + Pro pricing strategy and Stripe implementation plan where company-level subscriptions, full-access trials, checkout/webhook sync, billing settings, plan badges, and feature gates unlock clean PDFs, quote emails, and advanced branding. +- Technologies: Stripe, PostHog, Next.js route handlers/server actions, Drizzle, company subscription fields. +- Result: SaaS monetization architecture adapted from HoverNotes patterns while preserving Quest-specific company ownership and trial logic. +- Evidence: `docs/reference/stripe-pricing-subscriptions.md`; `docs/progress/hamza/CURRENT.md`; `src/app/pricing`; `src/app/app/settings/billing`; `src/app/api/stripe/webhook`. + +### Quote PDF And Email Delivery + +- Problem: Safari operators need professional, branded quote documents and a way to send them to travelers from the platform. +- What Hamza built: Quote PDF integration with preview/edit flows, page settings, R2-backed assets, quote email composer with Resend, merge tags, branded sender domains, logo controls, and PDF attachments. +- Technologies: PDF service integration, Cloudflare R2, Resend, Next.js server actions, typed PDF data transforms. +- Result: Operators can compose branded quote emails and deliver generated quote PDFs through Quest workflows. +- Evidence: `docs/progress/hamza/CURRENT.md`; `docs/progress/pdf-generation-plan.md`; `src/lib/pdf/transform-quote-to-pdf.ts`; `src/lib/email/*`; `quote-email-actions.ts`. + +### Data Management + +- Problem: Safari quote accuracy depends on structured destination, activity, hotel, room-rate, transport, company, and user data. +- What Hamza built: Admin and operator content-management areas plus seed/import/update scripts for domain data, including destination/activity/accommodation/transport content, images, themes, vehicles, staff, reviews, and company-specific settings. +- Technologies: Drizzle ORM, PostgreSQL, TypeScript scripts. +- Result: Maintainable data foundation for the platform and a path toward company-scoped content libraries on top of Quest defaults. +- Evidence: `quest-web-app/package.json` scripts; `README.md` admin panel list; `src/app/app/content-library`; `src/app/admin`; `docs/progress/hamza/CURRENT.md`. + +### Exchange Rates And Quote Economics + +- Problem: Uganda tour operators quote travelers in USD while many operating costs are in UGX. +- What Hamza built: USD/UGX exchange-rate settings, live/manual rate refresh, daily cron behavior, and transport fuel cost conversion into quote pricing. +- Technologies: Next.js, Drizzle/PostgreSQL, Open Exchange Rates needs confirmation, cron routes. +- Result: Quotes can show more accurate local-cost-to-client-price calculations for transport and fuel. +- Evidence: `docs/reference/exchange-rates.md`; `docs/progress/hamza/CURRENT.md`; `src/app/api/cron/update-exchange-rates`. + +### Release And Deployment Workflow + +- Problem: Quest needs a reliable shared QA and production release process. +- What Hamza built: Vercel-oriented deployment scripts and documented branch promotion workflow using `test`, release branches, and production promotion. +- Technologies: Vercel, shell scripts, Git workflow. +- Result: Clear handoff path for feature QA and production releases. +- Evidence: `quest-web-app/README.md` deployment and branch promotion sections. + +## Technical Stack + +- Frontend: Next.js 16, React 19, TypeScript, Tailwind CSS 4, shadcn/ui, Radix UI, dnd-kit, TipTap. +- Backend: Next.js App Router, API routes, server-side logic. +- Data: PostgreSQL, Supabase, Drizzle ORM, migrations and seed scripts. +- Auth: NextAuth v5, Google OAuth. +- Email/analytics/payments: Resend, PostHog, Stripe. +- Storage/media: Cloudflare R2, image upload/crop workflows, PDF preview screenshots. +- Deployment: Vercel, documented branch promotion workflow. + +## Resume Bullet Bank + +- Co-founded Quest and leads technology as CTO, leading a 3-person engineering/product team building a Next.js 16 safari quotation platform for Uganda-based tour operators. +- Built operator workflows for travel requests, tourists, itinerary planning, quote generation, PDF preview/export, quote email delivery, billing, analytics, and content libraries. +- Designed role-aware platform surfaces for Quest admins and tour operators using NextAuth v5, PostgreSQL/Supabase, and Drizzle ORM. +- Implemented structured data workflows for destinations, activities, hotels, room rates, transport, companies, and users to improve quote accuracy and operational maintainability. +- Designed Quest's Free + Pro launch model with company-level Stripe subscriptions, full-access trials, billing settings, plan badges, and output-level feature gates for clean PDFs/email/branding. +- Built quote PDF and email delivery workflows with R2-backed assets, Resend, branded sender domains, merge tags, page settings, preview regeneration, and PDF service integration. +- Established Vercel-based deployment and branch promotion workflow across feature branches, personal QA branches, shared QA, release branches, and production. + +## Site/Portfolio Angles + +- Useful as a portfolio case study for domain-specific B2B SaaS in travel/tourism. +- Strong angle: full-stack product engineering for a real operational platform, not a generic demo. +- Strong angle: data-heavy quote generation workflows with admin and operator surfaces. + +## Proof Links And Evidence + +- Local repo: `private-local-workspace/quest/quest-web-app` +- Repo README describes product, stack, user roles, admin panel, operator dashboard, and deployment workflow. + +## Needs Confirmation + +- Exact start month and whether to use `2026 - Present` or a specific month. +- Exact public role title: user stated CTO and co-founder. +- Launch status, customer/user scale, and any measurable business outcomes. +- Which features Hamza personally built versus inherited. diff --git a/career_evidence/roles/uganda-community-korea.md b/career_evidence/roles/uganda-community-korea.md new file mode 100644 index 0000000..61adb2d --- /dev/null +++ b/career_evidence/roles/uganda-community-korea.md @@ -0,0 +1,165 @@ +# Uganda Community In South Korea + +## Metadata + +- Evidence status: Verified from explicit user statement, local community workspace, and local PDF/PPT artifacts; public framing and acronym need confirmation +- Public name: Uganda Community in South Korea / Ugandans in South Korea Association needs confirmation +- Organization/client: USKA / UISKA / UCISK needs confirmation +- Role/title: Vice President +- Period: 2026 - Present needs confirmation +- Location: South Korea +- Source workspace: + - `private-local-workspace/uganda-community-in-korea` +- Existing public artifacts: Not yet integrated into site/resume +- Private/internal artifacts: + - `UISKA Core Values - Leadership Guide.pdf` + - `UISKA Core Values - Adoption and Practice Guide.pdf` + - `UISKA Core Values - Training Deck.pptx` + - `USKA_Embassy.pdf` + - `UGANDANS IN SOUTH KOREA ASSOCIATION proposal to Embassy.pdf` + - `USKA Partnership Strategy Map.pdf` + - `KAFCON 2026 - Team Uganda Matchday report.pdf` + - `REPORT TO THE USKA EXECUTIVE COMMITTEE.pdf` + - `ucisk-constitution (1).pdf` + +## One-Line Positioning + +Served in executive leadership for the Ugandan community in South Korea, helping formalize values, governance culture, diaspora partnership strategy, embassy-facing materials, and community engagement around sports, culture, welfare, and Uganda-Korea relations. + +## Context + +The Uganda community in South Korea is a diaspora association for Ugandans living in the Republic of Korea. Local materials describe the association as a platform for unity, welfare, cultural exchange, sports and entertainment, professional networking, and positive representation of Uganda in Korean society. + +The local evidence set spans older constitutional material under `Uganda Community in South Korea` / `UCISK`, newer association-facing materials under `Ugandans in South Korea Association` / `USKA`, and user-drafted core-values materials under `UISKA`. The final public acronym and organization naming should be confirmed before publishing. + +## Ownership + +- Executive/community leadership as Vice President, based on the user's explicit statement on 2026-06-05 and `USKA_Embassy.pdf` listing Hamza Kyamanywa as Vice President. +- Governance and leadership culture work: core values, values-vs-constitution explanation, practical adoption guide, and facilitation deck. +- Strategic communication and institutional storytelling for diaspora community work: embassy-facing proposals, association snapshot, partnership strategy, and community calendar. +- Support for community sports and cultural diplomacy narratives around KAFCON 2026 and Uganda's AFCON 2027 "Pamoja" moment. +- Leadership framing across respect, servant leadership, integrity, transparency, and accountability. + +## Major Workstreams + +### Core Values And Governance Culture + +- Problem: The leadership team did not clearly understand what core values are, why they matter, how they help, or how they differ from the constitution. +- What Hamza built: A leadership values package explaining core values, distinguishing them from the constitution, defining five values, and translating them into practical behaviors and decision questions. +- Technologies/process: PDF document creation, PowerPoint training deck, leadership facilitation structure, governance language, values adoption workflow. +- Result: A shareable values guide, an adoption/practice guide, and a 9-slide facilitation deck that can be used with the executive team. +- Evidence: + - `private-local-workspace/uganda-community-in-korea/UISKA Core Values - Leadership Guide.pdf` + - `private-local-workspace/uganda-community-in-korea/UISKA Core Values - Adoption and Practice Guide.pdf` + - `private-local-workspace/uganda-community-in-korea/UISKA Core Values - Training Deck.pptx` + +### Constitution-To-Culture Translation + +- Problem: Constitutional rules define offices, duties, elections, and governance procedures, but leaders still need shared behavioral standards for how authority should be used. +- What Hamza built: Messaging that explains the constitution as "what we are required to do" and values as "how we should do it." +- Technologies/process: Governance analysis, plain-language leadership writing, values checklist, executive team training material. +- Result: Leaders have a simple framework for connecting formal authority to respectful, humble, transparent, and accountable service. +- Evidence: + - `private-local-workspace/uganda-community-in-korea/ucisk-constitution (1).pdf` + - `private-local-workspace/uganda-community-in-korea/UISKA Core Values - Leadership Guide.pdf` + +### Embassy And Institutional Partnership Materials + +- Problem: The community needed credible materials for engaging the Embassy of Uganda, the Uganda Consulate in Busan, Korean institutions, and potential cultural/sports partners. +- What Hamza built or helped structure: Embassy-facing proposal language and strategic partnership framing around cultural diplomacy, diaspora development, sports diplomacy, educational/professional opportunities, and Uganda-Korea relations. +- Technologies/process: Policy-style writing, partnership mapping, calendar planning, diaspora strategy, stakeholder framing. +- Result: Local artifacts include a 2026 diaspora partnership proposal and a partnership strategy map covering academic institutions, Korean local authorities, African diplomatic missions, Korean corporate partners, cultural institutions, African community organizations, and international development/cultural agencies. +- Evidence: + - `private-local-workspace/uganda-community-in-korea/UGANDANS IN SOUTH KOREA ASSOCIATION proposal to Embassy.pdf` + - `private-local-workspace/uganda-community-in-korea/USKA Partnership Strategy Map.pdf` + - `private-local-workspace/uganda-community-in-korea/USKA_Embassy.pdf` + +### Community Snapshot And Data-Backed Association Profile + +- Problem: The association needed a concise institutional profile that could explain who it is, what it does, who leads it, and the status of Ugandans in Korea. +- What Hamza built or helped maintain: Association snapshot material with vision, mission, objectives, leadership structure, membership model, activity areas, calendar, and population/visa data for Ugandans in Korea. +- Technologies/process: Presentation material, demographic summary, Ministry of Justice/immigration data citation, community calendar planning. +- Result: `USKA_Embassy.pdf` provides a structured introduction to the association and cites 31 March 2026 population/visa figures. +- Evidence: + - `private-local-workspace/uganda-community-in-korea/USKA_Embassy.pdf` + +### KAFCON 2026 Sports Diplomacy And Community Engagement + +- Problem: Team Uganda's participation in KAFCON 2026 needed organization, reporting, and a narrative that connected sport to community cohesion and African/Ugandan visibility in Korea. +- What Hamza supported: Community leadership around Uganda's KAFCON participation, matchday reporting, executive reporting, and sports diplomacy framing. +- Technologies/process: Event reporting, match results tables, community appreciation messaging, sports/culture narrative. +- Result: Local artifacts document Uganda's participation in a 16-team tournament in Pyeongtaek City on 24 May 2026, with Uganda drawing all three Group D matches and using the event as a platform for unity and positive representation. +- Evidence: + - `private-local-workspace/uganda-community-in-korea/KAFCON 2026 - Team Uganda Matchday report.pdf` + - `private-local-workspace/uganda-community-in-korea/REPORT TO THE USKA EXECUTIVE COMMITTEE.pdf` + +## Metrics And Outcomes + +- Executive role: + - Status: Verified by explicit user statement and local `USKA_Embassy.pdf`. + - Evidence: User stated "i am the vp of the community"; `USKA_Embassy.pdf` lists Hamza Kyamanywa as Vice President. + - Public wording: "Vice President, Ugandans in South Korea Association" if acronym/name is confirmed. +- Association profile: + - Status: Verified from local artifact; public wording needs confirmation. + - Evidence: `USKA_Embassy.pdf`. + - Public wording: "Helped represent a registered 2026 Ugandan diaspora association in South Korea." Confirm whether to publish the registration number. +- Community size: + - Status: Verified from local artifact citing Korean Ministry of Justice/immigration data; should be rechecked before public use. + - Evidence: `USKA_Embassy.pdf` cites 586 registered Ugandans excluding short-term visas and 632 including short-term visas as of 31 March 2026. + - Public wording: "Supported leadership and communication for a Ugandan diaspora community of 600+ people in South Korea." Confirm before using. +- KAFCON event: + - Status: Verified from local reports. + - Evidence: KAFCON reports document 16 teams, 4 groups, 27 matches, Pyeongtaek City, 24 May 2026. + - Public wording: "Supported community representation and reporting for Uganda's participation in a 16-team African football tournament in Korea." +- Values/governance deliverables: + - Status: Verified local artifacts. + - Evidence: Core values PDF, adoption/practice PDF, training deck. + - Public wording: "Created governance and facilitation materials to help the executive team adopt shared leadership values." + +## Technical Stack + +- Strategy and governance: Core values, constitution-to-culture framing, executive meeting/adoption workflow. +- Communications: Embassy proposal, association snapshot, community reports, leadership training deck. +- Data and reporting: Population/visa data summaries, matchday reports, group standings, event outcomes. +- Tools/artifacts: PDF documents, PowerPoint deck, local evidence workspace, Markdown evidence bank. +- Domains: Diaspora leadership, cultural diplomacy, community welfare, sports diplomacy, partnership development, Uganda-Korea relations. + +## Resume Bullet Bank + +- Served as Vice President of the Ugandan community association in South Korea, supporting governance, leadership culture, community strategy, and institutional communication. +- Created a core-values leadership package defining respect, servant leadership, integrity, transparency, and accountability, with practical decision checks for executive service. +- Developed adoption and training materials to help community leaders translate values into visible behaviors, meeting rituals, onboarding practices, and accountability language. +- Helped frame diaspora partnership materials for embassy and institutional engagement across cultural diplomacy, sports programs, student/professional opportunities, and Uganda-Korea relations. +- Supported community communication and reporting around Team Uganda's participation in KAFCON 2026, connecting sport to unity, visibility, and positive diaspora representation. +- Helped organize association materials that explain mission, leadership structure, membership model, focus areas, community demographics, and activity calendar for external stakeholders. + +## Site/Portfolio Angles + +- Leadership case study: using plain-language governance materials to turn values into practice inside a volunteer executive team. +- Community systems case study: building structure, rituals, and communication around a diaspora association rather than only producing one-off events. +- Diplomacy/civic leadership angle: positioning diaspora work as cultural diplomacy, sports diplomacy, welfare, and Uganda-Korea relationship-building. +- Resume differentiator: shows non-software leadership, public service, strategy, communication, and stakeholder management alongside product/engineering work. +- Personal story angle: connects Hamza's technical/product work with real community leadership and service. + +## Proof Links And Evidence + +- `private-local-workspace/uganda-community-in-korea/UISKA Core Values - Leadership Guide.pdf` +- `private-local-workspace/uganda-community-in-korea/UISKA Core Values - Adoption and Practice Guide.pdf` +- `private-local-workspace/uganda-community-in-korea/UISKA Core Values - Training Deck.pptx` +- `private-local-workspace/uganda-community-in-korea/USKA_Embassy.pdf` +- `private-local-workspace/uganda-community-in-korea/UGANDANS IN SOUTH KOREA ASSOCIATION proposal to Embassy.pdf` +- `private-local-workspace/uganda-community-in-korea/USKA Partnership Strategy Map.pdf` +- `private-local-workspace/uganda-community-in-korea/KAFCON 2026 - Team Uganda Matchday report.pdf` +- `private-local-workspace/uganda-community-in-korea/REPORT TO THE USKA EXECUTIVE COMMITTEE.pdf` +- `private-local-workspace/uganda-community-in-korea/ucisk-constitution (1).pdf` + +## Needs Confirmation + +- Final public name and acronym: `USKA`, `UISKA`, `UCISK`, or another official form. +- Whether to use `Vice President` alone or a longer public title such as `Executive Committee Vice President`. +- Exact period: whether to use `2026 - Present`, election/appointment month, or another date range. +- Which artifacts Hamza personally authored vs reviewed, commissioned, or helped structure. +- Whether to publish the association's registration number from `USKA_Embassy.pdf`. +- Whether to publish community demographic figures such as 586/632 Ugandans in Korea, and whether those should be refreshed before publication. +- Whether to present this on the resume as leadership/community service, portfolio/case-study material, or both. +- Whether to mention KAFCON 2026 on the site/resume, and whether Hamza's personal participation as a squad member should be included. diff --git a/career_evidence/roles/xai-human-data.md b/career_evidence/roles/xai-human-data.md new file mode 100644 index 0000000..7ff0219 --- /dev/null +++ b/career_evidence/roles/xai-human-data.md @@ -0,0 +1,98 @@ +# xAI Human Data - Software Engineering + +## Metadata + +- Evidence status: Verified from contract, user statement, and public xAI role posting; internal task details confidential +- Public name: xAI Human Data - Software Engineering +- Organization/client: xAI / Remote Technology Services +- Role/title: Software Engineering Specialist, Human Data / AI Tutor - Software Engineering (Expert) +- Period: May 11, 2026 - Present +- Location: Remote from South Korea +- Source workspace: + - Private local workspace, not included in this repo. +- Contract evidence: + - Private contract PDF, not included in this repo. +- Public role evidence: + - `https://job-boards.greenhouse.io/xai/jobs/5063490007` +- Existing public artifacts: Not yet integrated into site/resume + +## One-Line Positioning + +Contracted on xAI's Human Data software engineering track to create and review high-quality coding data for training, benchmarking, and improving large language models. + +## Context + +The contract Statement of Work names the role as `AI Tutor - Software Engineering (Expert)` and defines the work as creating datasets for training, benchmarking, and advancing large language models. The public xAI Software Engineering Specialist posting describes related work: curating code examples, providing precise solutions/corrections, and evaluating/refining AI-generated code for efficiency, scalability, and reliability. + +Internal task details should be treated as confidential and not copied into public-facing materials. Public copy should stay at the level of software-engineering model evaluation, expert review, code-quality assessment, and coding-model improvement. + +## Ownership + +- Human expert review of AI-generated software-engineering work. +- Ranking model outputs across code quality, response quality, correctness, and overall task satisfaction. +- Reviewing AI-generated code and task context for correctness, maintainability, performance, security, and alignment with requirements. +- Using available tests, code behavior, and engineering judgment to support review decisions. +- Writing concise, evidence-grounded justifications that explain quality differences between model outputs. +- Screening tasks for cases where the available material is not reviewable. +- Working across multiple languages and domains, including Python, TypeScript/JavaScript, Java, Go, C/C++, Rust, infrastructure/distributed systems, databases, AI/ML/search, security, performance, and systems topics. + +## Major Workstreams + +### Coding Model Evaluation + +- Problem: AI coding systems need expert human review to identify which outputs best satisfy real software-engineering tasks. +- What Hamza does: Reviews model responses and generated code; ranks model outputs; and writes evidence-grounded justifications. +- Technologies/process: Coding-model evaluation, code review, task analysis, test-informed review, and concise technical writing. +- Result: High-quality expert labels and ranking rationales for coding-model training and benchmarking. +- Evidence: Private contract/workspace evidence and public xAI role posting. + +### Multi-Domain Software Judgment + +- Problem: Coding-model evaluation spans many ecosystems, not one narrow stack. +- What Hamza does: Applies software engineering judgment across languages and domains, including backend APIs, databases/ORMs, distributed systems, security, systems programming, performance, graphics, AI/ML, and tooling. +- Technologies/process: Python, TypeScript/JavaScript, Java, Go, C/C++, Rust, FastAPI, Kafka, OCPP, Hugging Face Transformers, Zed, delta-rs, CometBFT, Risc0, ray tracing, and related task domains. +- Result: Broader review coverage rooted in real engineering experience. +- Evidence: Private workspace evidence and public xAI role posting. + +### Evidence-Grounded Review Discipline + +- Problem: Model rankings are only useful if they are traceable to concrete code, response, and test evidence. +- What Hamza does: Writes structured review justifications that separate response quality, code behavior, test evidence, confidence, and reviewability concerns. +- Technologies/process: Review templates, code-anchor discipline, test-informed reasoning, and concise technical writing. +- Result: Review outputs are easier to audit and more useful for model improvement. +- Evidence: Private workspace evidence and public xAI role posting. + +## Technical Stack + +- Review domains: Python, TypeScript, JavaScript, Java, Go, C/C++, Rust, Shell, HTML/CSS. +- Engineering areas: AI/ML/search, databases/ORMs, distributed systems, security/privacy, performance/numerics, infrastructure/networking, graphics/media, language tooling. +- Workflow: coding-model evaluation, code review, ranking, test-informed reasoning, and final justifications. +- Caution: Do not publish internal task IDs, prompts, exports, rankings, private links, or detailed task content unless explicitly approved. + +## Resume Bullet Bank + +- Contracted with xAI's Human Data software engineering track as an AI Tutor - Software Engineering (Expert), creating and reviewing high-quality coding data for model training and benchmarking. +- Evaluated AI-generated code across correctness, maintainability, performance, security, test coverage, and alignment with task requirements. +- Ranked model outputs using evidence from prompts, generated code, responses, tests, and engineering judgment, then wrote concise justifications for training and benchmarking workflows. +- Reviewed coding tasks across Python, TypeScript/JavaScript, Java, Go, Rust, C/C++, databases, distributed systems, AI/ML, security, and performance domains. +- Applied production engineering judgment from SaaS, browser-extension, AI, payment, and infrastructure work to improve coding-model evaluation quality. + +## Site/Portfolio Angles + +- Strong signal for AI-model evaluation, coding-agent quality, and benchmark/training-data work. +- Useful for roles around AI coding systems, developer tools, model evaluation, data quality, or applied AI engineering. +- Should be described at a high level only; avoid internal platform/task details. + +## Proof Links And Evidence + +- Private contract PDF, not included in this repo. +- Private local workspace evidence, not included in this repo. +- `https://job-boards.greenhouse.io/xai/jobs/5063490007` + +## Needs Confirmation + +- Exact public title preference: `Software Engineering Specialist, Human Data`, `AI Tutor - Software Engineering (Expert)`, or both. +- Whether to name `xAI` directly on the public resume, given contract/confidentiality posture. +- Whether to mention `Remote Technology Services` as the contracting entity or omit it. +- Whether to use `May 2026 - Present` or `May 11, 2026 - Present`. +- How much of the coding-model evaluation workflow can be described publicly beyond the official role posting.