I build Python backends and AI-powered tools, with a focus on APIs that can be tested, deployed and maintained. My work includes FastAPI and Flask services, PostgreSQL, Docker, QA in staging, and LLM integrations through the Claude API and MCP.
I have contributed merged fixes to Hugging Face Transformers and MONAI. My background in Biomedical Engineering helps me approach software and data problems from more than one angle. I am enrolled in the BSc in Artificial Intelligence at UNIR.
Python · FastAPI · Flask · PostgreSQL · Docker · pytest · REST APIs · Claude API · MCP · LangChain
Full Stack Developer @ Nexeus Big Data (Sep 2024 – Mar 2025)
Built a gamified productivity and task management application with a Python backend and SQL.
Software Developer @ Software Praxis (Jan 2024 – Aug 2024)
Built the initial Flask HTTP API for the platform's Python backend and shipped changes to its WhatsApp bot. Contributed to an MCP server and tested releases in staging.
- Transformers #49157: removed an MPS autocast workaround across 20 files.
- MONAI #9125: confined
safe_evalname resolution to caller-supplied variables. - MONAI #8997: fixed an effective rank collapse score that could not reach its documented range after mean centering.
| Project | What it shows |
|---|---|
| Lumina Vault RAG | Local-first retrieval demo with FastAPI, PostgreSQL/pgvector, Redis, optional BGE embeddings, exact-quote citations, MCP tools and small labelled evaluations. |
| MedAPI | Patient management REST API with JWT authentication, role-based access control, PostgreSQL, Docker and pytest. |
| Steam Recommendation API | FastAPI service and ETL pipeline processing 100K+ Steam records for recommendations. |
| JobScout | Remote job matcher that uses the Claude API to rank listings against a candidate profile. |
| Heart Disease Analysis | R workflow comparing logistic regression, random forest and XGBoost on clinical data. |
| LinkedIn AI Recruitment Agent | Candidate sourcing prototype combining profile collection, LLM scoring and PostgreSQL. |