At Delivery Hero, one of the world's largest food delivery platforms, we run a decentralised data engineering org — 50+ teams, 70+ countries, 60+ PB data, 300k+ tables. Teams running autonomous pipelines, and domain engineers wanting to create data products rapidly enabling data driven business decisions. Complex setup, but initially all without dbt!

It won’t be a talk, but rather a live dialogue between two practitioners from opposite sides of that negotiation. One sat on the "central data platform team", responsible for standardisation, quality and governance. The other sat in a "domain team" owning their pipelines. We'll recreate the real friction — the "you can't do that" moments — and show how dbt gave both sides a shared language to work with.

How dbt forced the platform team to enable it: With its ease of use, small effort to migrate, suite of features like lineage and dbt tests. The features we've leaned on, the ones we're still exploring, and what we haven't figured out yet.