I'm Felipe, a full-stack software engineer based in Galway, Ireland. Currently completing my Higher Diploma in Computer Science @ CCT Dublin, mentoring on the Programming module, and building open-source AI dev tooling on the side.
More stuff about me
- I write code more than I write blog posts, but you can see what I'm shipping live at github.com/lipex360x.
- Long-form content in Portuguese: writing on dev.to/lipex360 (latest: validating environment variables in Node.js APIs) and a YouTube playlist on building a complete REST API end-to-end.
- Open to work in Ireland: eligible to work, no employer sponsorship required.
I spend most of my time these days on AI-assisted developer tooling. Not because AI is about to replace engineers, but because the engineers who can shape AI's output to match real production standards will outperform the ones who let it run unchecked. That belief runs through most of what I ship: keep the model honest, keep the test bar high, keep humans in the loop.
The headline projects:
- sourcemap-indexer: code-indexing tool for LLM-assisted dev workflows. Python, 629 tests at 100% coverage.
- dev-quality: linters, hooks and bootstrap scripts that hold AI-generated code to the same bar as anything else. Currently Python, pipeline ready for other languages.
- docx_builder: generates
.docxfiles (including my own CV) from a YAML description, so documents go through git like code. - github-toolkit: Bash dispatcher that scaffolds repos, labels, project boards, milestones and releases from JSON fixtures in a single command.
- cct-hdip-study: multilingual study materials for CCT classes, deployed at hdip-study-notes.vercel.app.
- web-view: project-agnostic Chrome/CDP toolkit on Playwright. Wait on state, never on time.
- Java and Spring Boot (AlgaWorks Backend Specialist, AlgaWorks Web Services and Microservices)
- Node.js and NestJS
- Python (FastAPI, LLM tooling)
- REST APIs, event-driven microservices, domain-driven design
- React and Next.js
- TypeScript, Tailwind CSS
- Vue.js
- AWS (Cloud Practitioner, AI Practitioner Early Adopter): SQS, SNS, Lambda, DynamoDB, S3
- Docker, Docker Compose, Nginx
- Vercel, Supabase
- Anthropic Claude API and Claude Code
- LLM workflows: code indexing, code review automation, document generation
- Custom Skills and Agents for repeatable tasks
- Jest, pytest, JUnit
- pre-commit,
dev-quality(own toolkit), 95%+ coverage targets - Code review as a first-class practice
| Language | Proficiency |
|---|---|
| Portuguese | Native |
| English | Professional working proficiency |
- Java internals and JVM tuning, via the CCT HDip curriculum
- Spring Boot 3 advanced patterns (AlgaWorks track)
- Building custom extensions for
pi-coding-agentin TypeScript, so I can compare different AI harnesses against Claude Code - Better prompting for code review and refactoring workflows
A few unusual bits about my setup that people sometimes ask about:
| Thing | Why |
|---|---|
| Shared "brain" directory synced across machines | Memory, skills and sub-agents are git-versioned and identical on every machine I touch |
| Telegram bridge to my coding agent | I can hand the agent tasks from my phone and get the results back asynchronously |
| Custom Skills + Agent definitions | Repeatable tasks (code review, document generation, Gmail filter management) live as small files |
docx_builder for my own CV |
This CV is generated from YAML and rebuilt on every change, so document edits go through git |