Graduate student in mathematics, machine learning, and computer science, building applied AI pipelines and open-source software. Curious, creative, and a hands-on generalist.
Why jsonlines everywhere? JSON is the first data format I've heard of and love using. Also, and maybe in the first place, it's a nearly perfect anagram of my real name.
From-scratch tensor operations, custom scikit-learn estimators, cross-validation harnesses, and proximal optimization routines developed alongside the École Polytechnique M2DS track.
Multi-agent B2B travel orchestrator built at the École Polytechnique AI Hackathon, combining OpenAI structured extraction, live travel quotes, multi-calendar constraint satisfaction, and disruption recovery.
End-to-end NLP classification pipeline using Hugging Face Transformers and FastAPI to match live news articles with social impact initiatives in real time.
Offline-first mobile cognitive aid and real-time clinical timekeeper for cardiopulmonary resuscitation (CPR) used by emergency medical teams (SAMU, SMUR), officially endorsed by the French Society of Emergency Medicine (SFMU).
Automated Python OCR pipeline for aeronautical maintenance documentation and custom TypeScript SharePoint Framework component connected to GitLab APIs at Safran Landing Systems.
Open-source mobile student life and grade-tracking client for French academic portals (Pronote, École Directe), scaling to over 10,000 downloads on iOS and Android.
Classes Préparatoires (MPSI/MP) TIPE research project performing automated grid colorimetry and temperature field extraction on thermal infrared imagery in Python.
Open-source JavaScript and TypeScript client libraries providing AES QR-code authentication deciphering and typed REST API wrappers for French educational platforms.
Machine learning theory, convex optimization, deep learning for computer vision, natural language processing & LLMs, audio/signal AI, and data stream processing.
Computer architecture, operating systems, distributed networks, statistical modeling, and continuous optimization (GPA 3.8/4.0). Finalist in the Concours Entreprendre Startup Competition (Project Ozy) and President of Club Xtreme.
Built and deployed a zero-shot semantic classification pipeline using Hugging Face Transformers with an asynchronous Python backend and embeddable web widget.
Automated technical documentation workflows using Python and OCR, and built custom TypeScript / SharePoint components connected to GitLab APIs.
Three-year intensive undergraduate preparation in abstract algebra, multivariable analysis, spectral theory, and physics.