MLOps can be defined as a set of practices that help you deploy and manage ML models in production environments. In this talk, I’ll first discuss what kind of issues come up when you move your ML workflow from development to production, and what approaches and design patterns can help. Then, I’ll talk about how you can operationalize these patterns using Google Cloud Platform as well as Kubeflow Pipelines.
Amy Unruh is a developer relations engineer for the Google Cloud Platform, where she focuses on machine learning and data analytics. Amy has a PhD in CS/AI, has worked at a variety of R&D labs and startups, and has written a book on App Engine.
