From PoCs to Large-Scale ML Operationalization: Covering the End-to-End Pipeline

Luxury department store Breuninger is one of the most successful fashion and lifestyle businesses in Germany today. With our considerable growth, the Data Platform Services team has faced the challenge to professionalize its services.

Over the last months, we have taken up the “MLOps Challenge”, i.e. moving away from a set of prototypical solutions operated by data scientists throughout the company to a MLOps platform solution associated with governance and a certain level of standardization.

In this talk, you will learn how the lifecycle of data – from ingestion to model output – feels like in what we see, if not as best, still as very good practice. You will learn how MLOps and ML Governance are handled at Breuninger and will gain insights into its operationalization.

Vorkenntnisse

• Fähigkeit, Code in SQL und Python zu lesen und zu verstehen

Lernziele

• Gründe für MLOps verstehen
  
• Qualitätskriterien für MLOps-Lösungen einschätzen können
  
• Hands-on-Erfahrungen in einem MLOps-Workflow End-to-End machen