Analytics and AI with PostgreSQL - Mission Possible!
summary
Claims a concrete stack and outcomes: EDB Postgres Lakehouse using object store + Parquet/Iceberg formats and separation of compute, PGAA to push tables to S3, PGFS/PGVector for storage and vectors, and AIDB to build retrievers and prompt flow inside Postgres. Live demos show a 6M‑row query hitting parquet files in ~12s, a vector upload flow (49 → 59 rows) and the vendor claim of up to 30× faster analytics at 5× lower cost; replication and GPU roadmap are discussed.
AI-written summary · about
chapters (10)
- 0:00PostgreSQL and AI analytics
- 3:20PostgreSQL strengths and limits
- 4:03EDB Lakehouse for real-time analytics
- 10:30Live demo: analytics with PostgreSQL
- 12:38AI models and neural networks
- 18:05Chatbot with PostgreSQL and vectors
- 25:34AI workflows with AIDB
- 27:00Live demo: AIDB vector search
- 31:13PostgreSQL AI and analytics roadmap
- 32:20Q&A: performance and editions
related talks
34mAnalytics and AI with PostgreSQL - Mission Possible!
50mFrom Transactions to Intelligence: Postgres for Open AI & Analyti
27mBuilding a PostgreSQL data warehouse
49mCan PostgreSQL Compete with Analytics Specialists?
21mAnalytics in Postgres –A decade in the making
37mRound Table Discussion
27mPutting AI in Production
44mIntegrating AI with Postgres...