From RAGs to riches: Build an AI document interrogation app in 30 mins
summary
Claims a working RAG app in 30 minutes and actually shows a runnable stack: Ragna (Python + REST), a Panel web UI, Chroma vector store (lens DB mentioned), and assistants like GPT 3.5, gp4, anthropic and Mosaic ML. Demo prepares embeddings, serves async queries, and builds a working RAG app in 30 minutes. A local ~4GB Hugging Face model runs on GPU and answers questions, the response objects include source chunks to check hallucinations, and the deployed web UI sometimes struggles under load but produces answers.
AI-written summary · about
chapters (7)
- 0:00RAG and LLM limits
- 7:29What RAG is and context fixes
- 14:38Ragna’s Python API setup
- 18:15Custom assistants in Ragna
- 20:40Ragna with local LLM
- 23:38Ragna’s design goals
- 26:08Q&A on embeddings
related talks
41mFrom RAGs to riches: Build an AI document interrogation app in 30 mins
02h 40mFrom RAGs to riches: Build an AI document inquiry web-app
01h 16mForget RAG Pipelines—Build Production Ready Agents in 15 Mins
42mTake Your RAG To The Max!
20mRetrieval Augmented Generation for AI Apps
44mThe RAG Really Ties the App Together
21mHow RAG is Redefining AI Capabilities
01h 13mBuilding Production-Ready AI Systems in 90 Minutes