Community Open Hours: Multimodal Search With PostgreSQL pgvector

In this episode, Denis and Franck are joined by Brett Hoyer, Senior Full-Stack Engineer at Yugabyte, to build a multimodal generative AI application using PostgreSQL pgvector.

We'll start by creating an application prototype that takes an image and text as input to generate a detailed image description. Next, we'll use an embedding model to convert these descriptions into multidimensional vector arrays. Finally, we'll demonstrate how to use the generated embeddings and Postgres pgvector to search for food recipes stored in a database.