Modern PostgreSQL for AI...
summary
Demonstrates building a retrieval-augmented chatbot over a 600-file codebase using PostgreSQL plus the open-source Lantern extension, Llama 3B for generation and Mistral for embeddings, storing vectors as real[] in tables and running similarity search via PG-based vector queries. Shows external serverless indexing to create large HNSW indexes with up to 100× more CPU than the DB, claims 3–9× price‑performance vs cloud rivals and reports the demo works end‑to‑end in about 15 minutes.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:09Building AI apps with PostgreSQL
- 4:23Demo: LLM chatbot setup
- 15:34Scaling the chatbot
- 27:22PostgreSQL extensions in cloud
- 33:44Underlying infrastructure
- 39:03Key takeaways summary
- 43:39Q&A session
related talks
27mPutting AI in Production
25mPostgres and the Artificial Intelligence Landscape
24mFrom Queries to Agents: The Next Era of Data Retrieval on PostgreSQL
01h 00mPostgreSQL and Artificial intelligence
37mRound Table Discussion
50mFrom Transactions to Intelligence: Postgres for Open AI & Analyti
35mPostgres in Artificial Intelligence
08mGenerative AI app development with Cloud SQL for PostgreSQL