Training AI on Your Own Data


Generative AI models are built on huge amounts of data. But not on your data. So how can you leverage your own proprietary source of truth with these "generic" models?

This session will explain how to ingest and leverage your own data in your own database to answer natural next questions. Using models with your own data can take several forms. Thist talk will focus on a technique using embeddings to do so.

We'll talk about generative AI models and theory, as well as how to put it all together for your own needs. Learning the basics of KNN / ANN, vectors, and vector search will help cement a mental model of what is going on behind the scenes.

Putting it all together we'll also show a practical implementation. Using a concrete example will show you to benefit from the power of AI without the expensive training and building of full-blown models.


Devs, DBAs, Architects, Analysts - all are welcome!

Ballroom F
Sunday, March 17, 2024 - 13:45 to 14:45