Case Study: Zalando achieves real-time personalization at scale with Hopsworks

Zalando delivers sub-10ms feature serving for 50 million customers with Hopsworks

Zalando, Europe's largest online fashion platform, faced challenges with fragmented data silos, inconsistent features, and high operational overhead in maintaining its machine learning infrastructure. These issues made it difficult to ensure low-latency feature access for real-time customer personalization across its 50 million customers. To address this, Zalando adopted the Hopsworks Feature Store to create a centralized platform.

By implementing Hopsworks, Zalando achieved sub-10ms latency for online feature serving, enabling real-time personalization like recommendations. The solution provided high availability with multi-availability zone replication and improved operational efficiency through Kubernetes-based autoscaling. Hopsworks helped centralize feature governance, reduce manual effort, and ensure predictable performance at scale.

Zalando