Unlocking Large-Scale Resilient Model Training with SageMaker HyperPod

In this hands-on workshop, participants will experience end-to-end model training and deployment using Amazon SageMaker HyperPod orchestrated with Amazon EKS — a purpose-built self-healing infrastructure that removes the undifferentiated heavy lifting of managing persistent clusters. You'll work with a live HyperPod EKS cluster to run distributed training across multiple GPU instances, test automatic fault recovery when hardware failures strike, monitor cluster performance with CloudWatch Container Insights, configure multi-team resource sharing with HyperPod Task Governance, and deploy a production inference endpoint. Participants will leave with practical experience across the full HyperPod lifecycle — from cluster operations and distributed training to resiliency, observability, governance, and serving — and a clear understanding of how HyperPod enables teams to focus on model development instead of infrastructure management.