Set Up an Open Source RAG in 30 Minutes
summary
Shows building a complete open-source RAG pipeline in minutes using LangChain for prompt construction and inference orchestration, OpenSearch 2.8 connectors to link an external LLM endpoint, and Agents 2.13 to coordinate retrieval then prompting. The demo stores 384‑dim embeddings with cosine similarity, vectorizes queries, retrieves documents and sends them plus context to a Mistral-based model endpoint (noting CPU inference is slow), and returns a numeric answer (a population delta ≈ 30,300) with a cost comparison to GPT-4 (~$0.09/1k tokens).
AI-written summary · about
chapters (7)
- 0:00Introduction
- 0:44LLM limits
- 4:34RAG principle
- 7:42Why open-source RAG
- 9:07RAG architecture with OpenSearch
- 12:02Step-by-step demo
- 21:23Conclusion & Q&A
related talks
34mImplementing an open-source RAG with OpenSearch
59mRAG Against the Machine
16mOpenRAG: An open-source stack for RAG
53mBuild RAG from Scratch
27mElevating AI Applications with OpenSearch's Flow Framework and RAG Tool
41mFrom RAGs to riches: Build an AI document interrogation app in 30 mins
01h 28mFrom RAG to Riches Building Enterprise Search Systems with LLMs
56mRAG: from dumb implementation to serious results