In this video, Abhishek, co-founder and CTO of True Foundry, shares insights on building a machine learning platform on Kubernetes. He discusses challenges such as handling diverse hardware, large Docker images, and slow iteration cycles. Abhishek also covers the importance of cost optimization, auto-scaling, and model caching. Learn how True Foundry addresses these issues by creating efficient abstractions and using tools like Argo CD, Keda, and Istio. Watch to understand how to streamline ML operations on Kubernetes.