Seattle Postgres: Vectors – Best practices for a nasty data type

Founded in 2010, the SEATTLE POSTGRESQL USERS GROUP (SEAPUG) gets together in person and we talk about technology and databases – especially our favorite database, POSTGRES! Nothing is too simple or too complicated. Come meet the people… from the newbie to expert – whether your job directly involves databases or not – all have a good time talking shop about technology over some food and drinks. ------------------------- Vectors are a centuries old, well-studied mathematical concept, yet they pose many challenges around efficient storage and retrieval in database systems. Applications requiring effective search techniques for vectors have advanced, with “retrieval-augmented generation” (RAG) becoming a key building technique. An extensible database like PostgreSQL can add vector search through an extension like pgvector. In this talk, we’ll review what vectors are, how they are used in applications, and what users are looking for in vector storage and search systems. We’ll then see how you can search for vector data in PostgreSQL, including looking at best practices for using pgvector by taking a deeper look at how pgvector implements different vector search techniques. We’ll also see where traditional databases methods are most effective for building RAG-driven apps. At the end of this talk, you’ll have a set of best practices you can use when designing applications that require vector search. Pizza is provided, and we also have a user group lending library of books & papers related to PostgreSQL and Databases. Speaker: Jonathan Katz is a Principal Product Manager – Technical on the Amazon Redshift team and is based in New York. He’s on the Core Team of PostgreSQL and is an active open source contributor, including to pgvector.