Applied AI · Automation · Python backend
I build Python backends, APIs and data pipelines, with a focus on applied AI: document processing, retrieval-augmented generation (RAG) and evaluation. My background combines software engineering with machine learning research.
Selected contributions merged into upstream projects:
- Presidio — analysis builder improvements and a NER model configuration fix.
- Presidio Research — configurable model evaluation and prediction, plus fixes for sample tag translation, CRF prediction tags and language handling.
- pdfly — page removal functionality for the PDF command-line toolkit.
| Project | What you can explore |
|---|---|
| Speech-based depression estimation | A research pipeline using DAIC-WOZ: speech preprocessing, model training and evaluation, with a Streamlit interface for data exploration. Dataset access is restricted; this is a research project, not a clinical tool. |
| azhacks | Async Python utilities for Azure Blob Storage, queues and tables. A small library focused on reusable storage operations. |
| Football for Burundi | A React landing page for the NEAR YOU charity football tournament in Madrid, held on 27 June 2026. |
- Backend and automation: Python, FastAPI, async programming, Azure and n8n.
- Applied AI: document processing, RAG, NLP and LLM evaluation.
- Research and data: PyTorch, speech analysis and Streamlit.
BSc in Data Science and Engineering and MSc in Machine Learning for Health, Universidad Carlos III de Madrid.
My public repositories include independent tools, research projects and open-source contributions. For professional experience and contact details, visit my portfolio.