Scaling Open Source ML: How Wolt Uses Kubernetes to Deliver Great Food to Millions

Discover how Wolt leverages Kubernetes to scale open-source machine learning for efficient food delivery in this 33-minute conference talk from KubeCon + CloudNativeCon. Explore the challenges of scaling ML infrastructure for over 12 million users and learn about Wolt's end-to-end MLOps platform built on Kubernetes. Gain insights into the integration of open-source frameworks like Flyte, MLFlow, and Seldon Core. Delve into topics such as forecasting supply and demand, restaurant recommendations, and delivery time predictions. Follow the machine learning lifecycle, deployment strategies, and the Wolt platform's ML flow. Examine complex inference graphs, integrations, and a live demo showcasing classification and the ML Flow UI. Understand how Wolt scales its ML platform to handle big ML models and utilizes Kafka and Snowflake. Conclude with a glimpse into future work and a Q&A session.