From vision research to product in prod.

I started in 2015 with applied computer vision and deep learning at a research lab in La Rochelle. My first IEEE papers were on B&W image colorization and face recognition in the visible+infrared domain.

After 4 years, I gradually moved to web development — first out of curiosity, then because I particularly enjoy the product rhythm: watching a design become a commit, become a feature used by real people in a matter of weeks.

I work at Okeenea on Evelity, a wayfinding app accessible to everyone, with a particular focus on the needs of visually impaired, deaf and mobility-impaired users.

I also contribute to Evelity Vision, the indoor navigation module powered by VPS: the user takes a photo, the system locates them inside the building. On the back-office side, I shipped the part that traces the trajectory of a 360° scan of the building on the floor plan and georeferences it.

Technically, I'm as comfortable with industrial Java/Spring as with data-heavy Python. On the front side it's React, TypeScript, Vite. I value tested code, constructive reviews, demanding CI pipelines — not as dogma, but because they save time over the long run.

Outside work, I do a lot of side-projects (see /en/projects), often as an excuse to learn a new technology or automate something tedious.

I walk my dog around the neighborhood, and that's where Klabi came from: a tiny app so my dog-owner neighbors and I can run into each other more easily. It's my side-project compass: write code to solve a real-life thing, not to look pretty on a résumé.