Data & AI Engineer
Freshly graduated engineer (Data & AI) - this portfolio walks you through my projects and my skills.
Scroll
Fresh out of ECE Paris, Data & AI track. I got my start at Renault Group, in data engineering, then at Wakam, in full stack development. I'm convinced hand-coding has less and less of a future, AI already automates a good share of it, and what really matters is knowing how to make the right calls: understanding the data, the architecture, the product. That's what's drawing me toward a role in data or AI, a sector with real staying power. Outside of client work, I build and host my own projects, down to the server they run on, mostly out of curiosity.
A few personal projects I build and host myself, plus some group projects from school.
The volleyball club's old site was outdated, badly organized and confusing. I volunteered to rebuild it from scratch, design to deployment.
I keep it updated all season long, results, photos, news, like the club's own little blog.
My graduation project, built with a team of six. Most of the group majored in finance and worked on the pricing and arbitrage models, I handled the site, front and back.
The most technical project of the bunch, and my first real taste of working with people who don't code.
I learned to solve the Rubik's Clock on YouTube. Once I had the method down I wanted to go fast, but beginner technique meant bad times. I fell down the speedcubing rabbit hole, algorithms, phases, techniques, and wanted a real timer for myself. So I built one.
A timer built for me, with the completion algorithms always at hand.
Picking a movie as a family is always the same struggle. The Tinder-style swipe idea was already floating around school and social media, I made it mine to practice, and to try out Claude's SuperPowers plugin along the way.
A simple project on the data side, more about testing a new tool than impressing anyone.
A two-person school project: a trip-itinerary recommender that adapts to budget and time available. My teammate handled the geospatial data (OSM, a graph of places), I built the recommendation model, the API and the interface.
My first real ML project: a Graph Attention Network trained to recommend places based on user context.
A group school project: a tool that reads a user's free-text answers and recommends a job profile and key skills. I focused on the semantic engine, SBERT sentence embeddings combined with TF-IDF to score closeness to each competency.
My first real dive into NLP: comparing free text against a skills database using sentence embeddings.
A group school project: analysis of the IMDB dataset (millions of movies and people), plus a Kafka pipeline that streams Wikipedia edits on popular movies and actors in real time to spot activity spikes.
My first real taste of Big Data: handling huge datasets and real-time streams with the same tools used in production.
I was running everything for free on Vercel, until curiosity won: I wanted my own domain names. With Claude by my side, I built my own infra, Cloudflare, Traefik, Docker, Coolify, on a Hetzner VPS. I provision, deploy and monitor it all solo, and this portfolio runs on the same stack.
Paris, France
French B2B2C digital insurer, Europe's leading embedded-insurance player, live in 32 countries.
Aubevoye, France
Renault Group's all-electric vehicle and software subsidiary, spun off in 2023.
Puteaux, France
French cleantech startup building smart collection points for food waste.
Boulogne-Billancourt, France
Global cybersecurity distributor, connecting vendors like Fortinet to resellers worldwide.