Snowflake OSS Meetup: AI up and down the stack
Join us for an evening of open source talks hosted by Snowflake on the theme of AI up and down the data stack. Talks will cover open source data services at all parts of the tech stack: how to manage and prep your data for AI ingest using Postgres, Iceberg and Spark, how to monitor your AI services using Trulens, and how to handle frontend deployment using Vercel and Streamlit. Join us for a fun-filled evening of talks from open source project maintainers from the Snowflake community and beyond!
6:00pm - Event starts, food and drink available
7:00 - 7:20pm - Holden Karau, Snowflake/Apache Spark PMC
Using Spark to Fight Health Insurance
In this talk Holden will walk you through how to use what she works on by day (Apache Spark) to to make AI models for social good. You'll learn a little bit about how American Healthcare "works", why peopple need to produce appeals, and how to use Apache Spark to produce training data for model fine tuning. We'll wrap up with a demo and some Q&A. For those sensitive to topics such as broken bones, we recommend you take the opportunity to grab a refreshment of your choice.
7:20 - 7:40pm - Break
7:40 - 8:00pm - Miles Richardson, EDB/WarehousePG maintainer
Bringing Postgres to the Lakehouse
In this talk, we'll introduce and demonstrate Postgres Analytics Accelerator (PGAA), an extension for reading and writing Iceberg (and Delta...) tables. We'll cover the technical details, including why we moved our focus from Delta to Iceberg, how we embedded DataFusion as a query engine into Postgres, and why we're now adding a Spark engine, too... But we'll also discuss customer requirements, what we're hearing from enterprise data teams, and how Iceberg allows us to "meet customers where they are," with an interoperable, incrementally adoptable path to modernization.
8:00pm - 8:20pm - JB Onofré, Dremio/Apache Iceberg maintainer
Talk TBD
8:20pm - 8:40pm - Speaker TBD