← Back to portfolio Live from private hardware
// Live RAG assistant

Ask my portfolio assistant

A small Llama 3 model running in Docker on private hardware. It answers from a little library of reference notes (retrieval-augmented generation) and cites what it used — so it stays grounded instead of guessing.

connecting…
Who's behind this site? How is this site built? What are the live demos? How do I get in touch?

Answers come from the reference notes when relevant, with sources shown. Runs on CPU, so replies take a few seconds.