Building Intelligent Applications with Graph-Based RAG on PostgreSQL

This session will show you how to revolutionize your applications using the Graph-Based RAG on PostgreSQL. The talk will show using PostgreSQL with vector and graph data to enable natural language queries for highly accurate results that will power your RAG application.

Vector data will be explained as a tool that enhances PostgreSQL's ability to manage information for GenAI apps. The session explores how to integrate graph, and vector data using azure_ai, pg_diskann, age and pgvector extensions. As well as advanced techniques to improve accuracy of results and improve vector search performance at scale.

A hands-on demonstration will show how to set up PostgreSQL to handle these types of queries, store vector data, and create queries that respond to natural language inputs. A practical example will involve building a legal copilot to help lawyers find relevant legal cases during research.

Participants will leave the session equipped to apply similar graph, indexing and vector features in their own applications, making them more intuitive, user-friendly and scalable.