On Demand

The Emerging Blueprint for Agentic Apps

Evolved stack, still Postgres
Reynold Xin smiling, data analytics diagram.

Why Postgres is the foundation for a new data processing architecture

Providing your agent with live application state, historical data and a unified security model means stitching together infrastructure with pipelines you need to monitor and fix.

What if you could skip the data transfer altogether? Lake Transactional/Analytical Processing (LTAP) lets transactions and analytics read from the same copy of data by storing it in open formats on standard Postgres.

Join Reynold Xin, Databricks Co-founder and Chief Architect, for an overview of this new architecture, a demo and benchmark. We’ll cover:

  • How LTAP brings agent memory, application state and historical data under the same permissions
  • Removing sync intervals as a bottleneck for the agent memory loop while maintaining performance and availability
  • Compatibility with your drivers, extensions and ORMs by running on standard Postgres

Speakers

Smiling Asian man in gray jacket.

Reynold Xin

Co-founder and Chief Architect

Databricks

Smiling man with glasses

Jonathan Katz

Sr. Staff Product Manager and PostgreSQL Core Team

Databricks

Hristo Stoyanov

Hristo Stoyanov

Sr. Software Engineer

Databricks

Agenda