About

I’m Alexandre Girard Davila. For twenty years I’ve built software that has to keep working when it matters — for Airbus, SNCF, Cartier, and Tate — across computer vision, edge computing, and real-time data.

My current focus is narrower and sharper: agentic AI that works in the demo and breaks in production. I find the failure modes, score them by how badly they’ll bite, and put them in order so your team can fix the things that actually matter.

What I’ve done

• Airbus — production computer-vision pipelines on the factory floor.
• SNCF — video metadata extraction from railway footage for predictive maintenance.
• Cartier (via JoliGen) — AI-powered generative art for luxury retail experiences.
• Tate Britain — Recognition, winner of the IK Prize 2016, an installation comparing photojournalism with the British art collection, built with Microsoft and Fabrica.

How I work

I’ve been the engineer, the lead, and the person who has to explain why the thing broke at 2 a.m. I’ve shipped Fortune-500 enterprise platforms and open-source tools that other teams adopt (Parasol, raspi-kiosk-server, DeepDetect JS, React Bounding Box). I care about the boring parts — observability, failure handling, the runbook — because that’s where “demo” becomes “production.”

I’m independent now. When you bring me a problem, you get my judgment directly, with no account manager, no discovery-call theatre, no discovery-to-discovery ping-pong. If I’m not the right person, I’ll tell you and point you to someone who is.

In short: two decades of shipping systems that couldn’t fail, applied to the specific problem of making AI behave the same way.