How Databricks uses Databricks
Overview
| Experience | Virtual |
|---|
Ever wonder how Databricks operates its own enterprise lakehouse, where all employees and all teams inside use data and AI to solve problems and guide our decisions? Bruce Wong, head of data platforms, will talk about how his team leverages Databricks itself at a massive scale and value.
Session Speakers
Ari Kaplan
/Global Head of Evangelism
Databricks
Bruce Wong
/Sr. Director, Data Platform
Databricks
Holly Smith
/DevRel
Databricks
Full Summary
Inside Databricks on Databricks: How the Company Runs on Its Own Platform
Databricks operates as customer zero for its own technology, running core business functions on the Databricks Platform and letting employees interact with data directly. The approach is cultural as much as technical, built on intellectual honesty and a bias to interrogate real data instead of relying on intuition or static dashboards.
FAQ
The internal platform team adopts preview and pre-release features against real production data before external customers use them. That early usage generates direct feedback that helps refine capabilities and validate product-market fit prior to general availability.
Teams use Genie to ask natural-language questions of their own data. Facilities uses badge data for capacity and planning, and HR analyzes culture survey responses interactively rather than relying on a single sentiment score.
Agentic AI adoption accelerated over the last six months, and employees now spin up agents to run recurring and continuous tasks. The internal environment hosts over 100,000 agents that interact with Databricks systems, far exceeding the human workforce.
The team built admin agents to govern other agents. These handle areas like cost management, administration, and data governance, making oversight practical at machine scale instead of relying on linear human intervention.
It sits within engineering by intent. Proximity to product management and technical leadership enables fast feedback loops and lets the team pressure-test features against production workloads early.