Jérémy F.

ML Engineer / AI Engineer full-stack, I design and deliver end-to-end AI products, from model to UI.

My background combinesmathematical training (École des Ponts ParisTech), 4+ years inML/Data in fintech and banking (Qonto, BNP Paribas), andfull-stack expertise (TypeScript, Next.js, React, FastAPI) that allows me to deliver complete products, not just notebooks.

Recently, I built Sajedition, a platform that provides students with a co-learner to better understand their courses, powered by a RAG system to minimize hallucinations, with a FastAPI backend, TypeScript/Next.js frontend, deployed on Vercel and a PostgreSQL DB.

I also created Vibetrip, a travel search platform powered by a multi-agent system with Deep Research (CrewAI), with a FastAPI backend, TypeScript/Next.js frontend, deployed on AWS with Terraform and PostgreSQL.

At Qonto, I deployed ML models in production with < 50ms latency and explored GNNs and the latest AI tools for the team. I've been using Claude Code and Codex for a long time, which allows me to deliver much faster today.

→ Complete AI Products : applications with LLMs, autonomous agents, RAG, chatbots — from prompt engineering to production deployment

→ Multi-agent Systems : orchestration with CrewAI, OpenAI, API and tool integration

→ ML in Production : performant models deployed on AWS with monitoring, robust MLOps pipelines

→ Full-stack : React/Next.js/TypeScript frontend + FastAPI/Python backend + AWS/Terraform/Docker/Kubernetes infra

→ Rapid Prototyping : I work with Claude Code, Cursor, and Codex to deliver quickly without sacrificing quality

Stack : Python, TypeScript, PyTorch, OpenAI, CrewAI, FastAPI, React, PostgreSQL, AWS

Available remotely or hybrid in Paris.