Surviving the 100-Team Shift: Architecting Accountability at Scale
Migrating ~100 autonomous teams and dozens of data pipelines to another technological stack while transitioning to a new event-tracking platform is typically a full-year undertaking. At Zalando, a small, dedicated platform team executed this shift in just three months by combining AI-driven automation with a product-led approach to data ownership.
In this talk, I will share the architectural and organizational blueprint behind this transition. We will explore how treating migration as a product enabled domain teams through flexible adoption paths and how we resolved the “orphan pipeline” dilemma without assigning blame. We will dive into the technical mechanics: leveraging multi-model AI, executing safe deprecation through structured scream tests, and cutting compute costs by 52% while eliminating cross-cloud egress fees.
Attendees will walk away with practical, battle-tested patterns for scaling data ownership, empowering autonomous teams, and executing large-scale migrations without business disruption.