Observatory — Desbouis, Gabriel — AI Engineer & AI Architect
AI Engineer and AI Architect. I build AI systems that actually run — not just in a demo. I learned the craft from both sides: the technical one (data science, ML, architecture) and the business one, because a model that doesn't create value is just a nice prototype. What I care about is the hardest part: getting from pilot to production.
Survey of works in progress
The biggest build: a full ecosystem — geographic & job-offer data pipelines, AI core, monitoring dashboards, agents. Designed end-to-end, from data collection to production.
Fullstack monitoring app: automatic collection, LLM classification and translation, structured newsletter and LinkedIn post composition.
Owner-assistance platform: centralizing construction-project information and helping non-expert decision-makers steer.
iOS app in public beta: building and keeping a schedule, simply. Distributed on TestFlight.
I got into AI through the technical door, and I stayed for what it lets you build. My background is hybrid: a computer engineering degree, a specialized master's in AI, but also an entrepreneurship degree and a few business games won along the way. The result: I think as much in 'how does it work' as in 'what does it return'.
Today a data science & ML engineer with an AI Architect hat, I design the architecture of AI solutions and take them to production. I like projects where you have to connect both ends: a model that works AND real value for the business — that famous PoC-to-production gap, where so many AI projects get stuck.
On the side, I explain: university tutoring, an AI newsletter. Making complex things clear without betraying them is the same discipline as putting them in production.
GabrielCorrespondence — reply within 48 h
An AI project to take to production, an idea to challenge, an opportunity? Write to me, I reply fast.
desbouis.lgabriel@gmail.com