I'm a Software Engineer working mostly on AI systems.
I started in data science, so I'm comfortable on both sides of the boundary: models and evaluation, but also architecture, testing and production software.
My work has often taken me outside my initial area of expertise, and I'm comfortable stepping into unfamiliar areas when the problem requires it.
I helped turn an internal RAG prototype into a production service used by 500+ people across three subsidiaries.
I built its automated evaluation pipeline around a business-curated Q&A dataset and LLM-as-a-judge, and worked on improving answer quality up to an 80% correct-answer rate. I also worked on the API, indexing pipelines, monitoring and reliability.
I was a lead maintainer of this open-source multiplayer game remake from 2022 to 2025.
I drove architecture, technical and product decisions, reviewed and mentored 20+ contributors, and helped grow the project to 2,500+ users.
Archéologie logicielle : reconstruire un jeu sans son code source
I've contributed to open source projects including uv, pytest, Haystack and Hatch.