Clara MacAvoy

Running a data platform on one cloud is standard — running it across AWS, Azure, and GCP simultaneously is advanced. But running it across 100+ distinct regions while maintaining a unified experience? That is the Databricks Data Platform team's daily reality. In this session, we reveal the architecture of our massive internal global mesh. We will dive into the challenges of data gravity, cross-region replication, and unified identity at a scale few companies face. Learn how we abstract away the chaos of 100+ regions to provide a seamless, singular "Databricks experience" to our internal developers — proving that you really can build a true multi-cloud lakehouse without getting locked in.

Struggling with runaway cloud costs as your organization grows? Join us for an inside look at how Databricks’ own Data Platform team tackled escalating spend in some of the world’s largest workspaces — saving millions of dollars without sacrificing performance or user experience. We’ll share how we harnessed powerful features like System Tables, Workflows, Unity Catalog, and Photon to monitor and optimize resource usage, all while using data-driven decisions to improve efficiency and ensure we invest in the areas that truly drive business impact. You’ll hear about the real-world challenges we faced balancing governance with velocity and discover the custom tooling and best practices we developed to keep costs in check. By the end of this session, you’ll walk away with a proven roadmap for leveraging Databricks to control cloud spend at scale.