LEGO bricks work because they are reusable by design. Your ETL should be too. Every new source should snap into the same framework, not require a custom-built pipeline from scratch.
In this half-day workshop, we will build a metadata-driven ingestion architecture in Microsoft Fabric from the ground up, using real LEGO pricing data as our working dataset. You will learn how to design a control table that acts as the single source of truth for what gets ingested, when, and how. From there, we will translate that metadata into notebook execution DAGs that handle dependencies, parallelism, and reruns without hardcoding a thing.
Along the way, we will work through the decisions that separate a pattern that holds up from one that collapses the first time requirements change. How granular should your control table be. Where does parameterization end and configuration begin. How do you structure notebooks so adding a new source is a metadata change, not a code change. We will mix walkthrough demos with hands-on exercises so you leave having actually built pieces of the framework, not just watched someone else do it.
This workshop is designed for data engineers who are newer to Fabric and want a blueprint they can take home and extend. By the end of the session, you will have a working reference implementation, a mental model for how the pieces fit together, and the confidence to apply the same pattern to your own sources the Monday after the event.
Bring a laptop. Bring your messiest ingestion problem. We will snap it together brick by brick.