Curious and fast learner, especially interested in new AI tools like Claude Code and how they work. I’m comfortable under pressure, and alongside my studies I practice boxing at a high level in France. I did a lot of freelance mission. My last one for the PIMA county in ATL where I led the development of an end-to-end ML detection pipeline from scratch. I started with limited infrastructure and DevOps experience, learned on the job, and successfully delivered a system that is now in testing with real clients. The project focuses on road sign detection from videos, followed by classification and condition assessment. I built a Kotlin mobile app to record and upload large videos (tens of gigabytes), which was technically challenging. On the backend, I developed a full Python API using FastAPI, with JWT-based authentication and a SQLite database. Redis queues handle uploads, and Celery workers process jobs. Each upload triggers an ML pipeline developed with PhD collaborators, running on dedicated AWS EC2 instances. Videos are stored on AWS S3. I also built a lightweight web dashboard (FastAPI) to manage uploads, visualize and download results, and filter by user, date, or organization. 
…Curious and fast learner, especially interested in new AI tools like Claude Code and how they work. I’m comfortable under pressure, and alongside my studies I practice boxing at a high level in France. I did a lot of freelance mission. My last one for the PIMA county in ATL where I led the development of an end-to-end ML detection pipeline from scratch. I started with limited infrastructure and DevOps experience, learned on the job, and successfully delivered a system that is now in testing with real clients. The project focuses on road sign detection from videos, followed by classification and condition assessment. I built a Kotlin mobile app to record and upload large videos (tens of gigabytes), which was technically challenging. On the backend, I developed a full Python API using FastAPI, with JWT-based authentication and a SQLite database. Redis queues handle uploads, and Celery workers process jobs. Each upload triggers an ML pipeline developed with PhD collaborators, running on dedicated AWS EC2 instances. Videos are stored on AWS S3. I also built a lightweight web dashboard (FastAPI) to manage uploads, visualize and download results, and filter by user, date, or organization. 
WWWW…