Data engineering is entering a new phase – going beyond AI that assists, towards autonomous agents that can execute complex, long-running work while humans focus on higher-value decisions.

In this session, we’ll introduce “Project Osmos”, Microsoft Fabric’s agentic data engineering experience. We’ll walk through a realistic end-to-end scenario from an ambiguous prompt to a complete, validated project. You’ll see how “Project Osmos” autonomously takes a complex task, inspects the workspace to understand the code and data layout, generates and updates notebooks and other Fabric artifacts, and validates the result so teams can trust what was built.

We’ll cover how data engineering work itself is being reimagined inside Fabric: long-running tasks, sub-agent orchestration, self-validation, guardrails, trust, and why human supervision will continue to play a critical role in the process.