Presented By
O’Reilly + Cloudera
Make Data Work
29 April–2 May 2019
London, UK
Please log in

Insights from engineering Europe's largest marketing platform for fashion

Dirk Petzoldt (Zalando SE)
12:0512:45 Wednesday, 1 May 2019
Case studies, Strata Business Summit
Location: Capital Suite 12
Average rating: ****.
(4.18, 11 ratings)

Who is this presentation for?

  • Tech leaders, engineers, developers, and data scientists



What you'll learn

  • Understand the biggest data challenges in online marketing
  • Learn how to build a big data team and architecture from scratch while steering a 9-digit marketing budget with machine learning and orchestrating a broad range of big data technologies
  • Explore lessons learned from a real project


In early 2018, Zalando—Europe’s leading online fashion platform—embarked on a journey to revolutionize its paid online marketing by building a scalable personalized machine learning–based marketing platform.

Dirk Petzoldt covers the overall architecture and data flow and reflects on lessons learned along the way. The solution leverages Spark, PostgreSQL, Scala, Java, Python, AWS, Docker, Kafka, Kubernetes, TensorFlow, Scalyr, SageMaker, and Databricks, among others.

Topics include:

  • Processing over 400 million events daily in batch and real time
  • A/B testing and attribution
  • Cross-device graph creation
  • Budget and ROI forecasting
  • Bidding and steering algorithms
  • Integration with Facebook and Google
  • Image mining
Photo of Dirk Petzoldt

Dirk Petzoldt

Zalando SE

Dirk Petzoldt is a head of engineering and data science at Zalando, Europe’s leading fashion platform. Trained as a data scientist, he enables his five development teams to revolutionize online marketing steering in a fully automated, ROI-driven, personalized way. In his spare time, Dirk is hacking functional Scala and reading through O’Reilly’s online library, 10 books at a time.