RAGtime with Postgres: AI Power with pgvector and Retrieval-Augmented Generation
327 | Fri 01 Aug 5:30 p.m.–6:15 p.m.
Presented by
-
Jimmy Angelakos
https://vyruss.org/computing
Jimmy Angelakos is a Systems and Database Architect and recognized PostgreSQL expert who has worked with, and contributed to, Open-Source tools for 25+ years. He is passionate about participating in the community, a Contributor to the PostgreSQL project, and an active member of PostgreSQL Europe and US. Jimmy is a regular speaker at conferences and events, sharing his insights with the community. Author of PostgreSQL Mistakes and How to Avoid Them, co-author of PostgreSQL 16 Administration Cookbook.
Mastodon: https://fosstodon.org/@vyruss
Bluesky: https://bsky.app/profile/vyruss.org
Jimmy Angelakos
https://vyruss.org/computing
Abstract
Retrieval-Augmented Generation (RAG) is a powerful paradigm in application development with AI. In this talk, we'll demonstrate how to leverage PostgreSQL with pgvector to combine the strengths of vector similarity search with Large Language Models (LLMs).
As the speaker is a Postgres nerd (not an AI expert), we'll explain in simple terms how to dip your toes into AI while leveraging our favorite database -- from the perspective of a database person learning to work with these new tools.
We'll walk through:
- How to use pgvector to store and search vector embeddings (and what those are)
- How to connect these capabilities with AI LLMs to build intelligent applications.
- Some practical tips for implementation, including configuration, indexing strategies, and scaling considerations
- How to reduce dependency on expensive external AI services by using open-source models while maintaining control over costs and infrastructure
To demonstrate these concepts in action, we'll look at a real-world example of building a developer assistance system that helps teams understand their codebase.
Retrieval-Augmented Generation (RAG) is a powerful paradigm in application development with AI. In this talk, we'll demonstrate how to leverage PostgreSQL with pgvector to combine the strengths of vector similarity search with Large Language Models (LLMs). As the speaker is a Postgres nerd (not an AI expert), we'll explain in simple terms how to dip your toes into AI while leveraging our favorite database -- from the perspective of a database person learning to work with these new tools. We'll walk through: - How to use pgvector to store and search vector embeddings (and what those are) - How to connect these capabilities with AI LLMs to build intelligent applications. - Some practical tips for implementation, including configuration, indexing strategies, and scaling considerations - How to reduce dependency on expensive external AI services by using open-source models while maintaining control over costs and infrastructure To demonstrate these concepts in action, we'll look at a real-world example of building a developer assistance system that helps teams understand their codebase.