MLOps at Snapchat - Continuous Machine Learning with Kubeflow & Spinnaker

Explore a conference talk on implementing MLOps practices at Snapchat using Kubeflow and Spinnaker. Discover how Snap Inc. transformed their manual script-driven process into a robust, automated machine learning pipeline for computer vision applications. Learn about the challenges of integrating and operating ML systems in production, and how applying DevOps principles to machine learning (MLOps) can help navigate the entire ML lifecycle. Gain insights into leveraging Kubernetes, Kubeflow pipelines, and Spinnaker to achieve continuous integration, continuous delivery, and continuous training for complex ML systems. Understand the unique differences between traditional software systems and ML systems, and how to address them in a production environment.

MLOps at Snapchat: Continuous Machine Learning with Kubeflow & Spinnaker - Kevin Dela Rosa, Snap Inc