From Livestream Chaos to Production‑Ready AI: RAG + Observability with CHARLIE

Speakers

AI features are easy to prototype and hard to trust. In this session, we’ll walk through a real example: CHARLIE, the chatbot that runs the csharpfritz channel livestream. We’ll start with his baseline architecture, then layer in retrieval‑augmented generation to give him real knowledge about commands, FAQs, and stream context.

Once CHARLIE can answer questions reliably, we’ll turn on Progress AI Observability to expose the full pipeline - retrieval calls, prompt construction, model outputs, latency, cost, and safety signals. You’ll see how observability transforms debugging from guesswork into engineering, and how these patterns apply to any developer building AI into their apps.

An unhandled error has occurred. Reload 🗙