Antoine Caille

I'm an AI Product Engineer. I design LLM agents, build the products around them, and I also do full-stack and data work when that's what the project needs.

At FuzzingLabs, I lead product and own the agentic layer of an offensive-security platform. Google ADK for agent orchestration, Cognee for memory and knowledge graphs, MCP and A2A for integrations, Python services over APIs, Next.js on the frontend, Kubernetes for deployment.

Before that: two years at Ekimetrics on marketing mix modeling for enterprise clients in banking, insurance, and QSR, then a short stint at LittleBigCode on a forecasting POC.

I also guest-lecture on LLM theory and architecture at engineering master's programs when the schedule allows.

I've shipped production LLM agents, not just prototypes. I've sat in client rooms explaining models to people who don't code. I can own a feature from Postgres schema to the Next.js component, and I've done it often enough to not break things.

- LLM agents with real memory (Google ADK, Cognee), from design to prod

- RAG pipelines, including graph-RAG when your data calls for it

- MCP / A2A integrations into an existing stack

- Next.js frontends for AI products — streaming, real-time, state sync

- Full-stack feature work when you need an extra senior pair of hands

- Data analysis with real business output, not a 40-slide deck

- Technical scoping before you commit budget

Full remote, French or English.

Send me your brief. I'll tell you if I'm the right person or not.