Boost LLMs with Event Streaming and Retrieval Augmented Generation
summary
Claim: Retrieval-Augmented Generation (RAG) cuts hallucinations and is far cheaper than re‑training foundation models while letting LLMs use up‑to‑date data. Implements event-driven pipelines with LangStream connecting Kafka/Pulsar sources to vector stores like Astra DB, Pinecone or Milvus, using ANN for nearest‑neighbor embeddings and declarative YAML pipelines. Demo uses Docker plus an OpenAI key and LangChain/LlamaIndex connectors; one notable metric cited is 1 million users in 5 days for GPT‑3.5.
AI-written summary · about
chapters (7)
- 0:00Introduction and agenda
- 2:15Drawbacks of LLMs and improvements
- 12:51Introduction to RAG pattern
- 23:27What is RAG and how it works
- 26:18Introducing LStream library
- 30:07Demo and quick start
- 33:36Resources and closing
related talks
01h 33mRAG How To Talk To Your Data
51mEvent Driven AI: Supercharging ChatGPT with RAG & LangStream
53mFrom RAG to Riches: Secrets of Data-Augmented LLMs
48mGoing Beyond Prompt Engineering
42mTo RAG or not to RAG
48mCrafting your own RAG system: Leveraging 30+ LLMs for enhanced performance
18mLLMs and GenAI: A Guide to Building Your Own RAG Model for Document Interaction
48mAn Introduction to Retrieval Augmented Generation