Pixeltable: A DBMS for Multimodal AI Applications (Marcel Kornacker)
- Speaker:
- Marcel Kornacker
- Date:
- Mon Mar 30, 2026 @ 04:30pm EDT
- Date:
- Mon Mar 30, 2026
- Time:
- 04:30pm EDT
- Location:
- https://cmu.zoom.us/j/99830697483?pwd=RLKiHNDLPvOyoCMHSBGmWfbba4ZAb4.1Zoom
- Title:
- Pixeltable: A DBMS for Multimodal AI Applications
- System:
- Pixeltable
- Video:
- YouTube
Talk Info:
Pixeltable is a novel OLTP DBMS specifically designed for multimodal AI applications. It goes beyond the traditional roles of a DBMS — data storage and retrieval — by incorporating data generation into the table schema. Pixeltable natively supports multimodal data types (e.g., video, images, audio, documents) in the context of traditional multi-user transaction semantics. Unlike analytics-focused databases, Pixeltable also allows the user to express data generation declaratively, via computed columns and user-defined functions. This makes database technology relevant again for the GenAI age. Pixeltable is fully open-source and available as a pip-installable Python package; it utilizes Postgres for local structured data.
This talk is part of the PostgreSQL vs. The World Seminar Series.
Bio:
Marcel is the CTO and co-founder of Pixeltable, the founder of Apache Impala, and the co-creator of the Parquet file format. He holds a PhD from UC Berkeley.