Case Study: Otto achieves unified data platform control and faster pipeline development with Dagster Labs

OTTO unifies 6 data teams with Dagster to replace YAML hell

OTTO, one of Germany's largest e-commerce companies, faced significant challenges with its fragmented data pipeline infrastructure. Different teams were using tools like Argo Workflows and Apache Airflow, leading to what they called "YAML hell," a lack of local testing, and complex ML orchestration issues that hindered their ability to respond to market changes. This patchwork of legacy orchestrators created operational overhead and limited collaboration.

By implementing Dagster, Otto unified six data teams onto a single platform. The solution provided Python-native development to eliminate YAML configuration, local testing capabilities for faster debugging, and a data-centric architecture with reusable asset factories. As a result, Dagster helped Otto improve pipeline stability, reduce failure resolution time, foster better cross-team collaboration, and create a foundation for broader organizational adoption of their data platform.

OTTO