Building Agentic Business Applications on Databricks with Lakebase, MCP and Marvis AI

Many organizations use Databricks for analytics and AI, but still build business applications on separate stacks, disconnected from the lakehouse. We share how we use Databricks as a full application platform, combining analytics, operational data and AI into production-grade systems. We will show how we built DNA Port and Marvis AI using Databricks Apps, Lakebase, Lakeflow Sync Tables and MLflow for governed model lifecycle management. We will also demonstrate how MCP securely exposes trusted lakehouse data to AI copilots. Through real-world demos, we show how user actions write to Lakebase, trigger AI workflows and surface insights in real time. Today, this platform supports 200+ vessels and 300+ on-shore employees, reducing manual reporting by 40% and improving application time to market by 50%. Attendees will learn practical architecture patterns, governance with Unity Catalog, and lessons learned in building scalable, AI-powered enterprise applications on Databricks. Talk By: Simon Fassot, Head of Data and AI, Hafnia ; Connect with us: Website: https://databricks.com X: / databricks LinkedIn: / databricks Instagram: / databricksinc Facebook: / databricksinc