Chance Lee

In this builders’ session, explore how to extend generative AI workloads using the capabilities of Amazon ECS with AWS Fargate and Retrieval Augmented Generation (RAG).

Karpenter, an open source CNCF project, is a flexible, high-performance Kubernetes cluster auto-scaler built with AWS. It helps improve application availability and cluster efficiency by rapidly launching right-sized compute resources in response to changing application load.

Optimizing for cost is a key consideration when designing or improving workloads. In this builders’ session, explore how to gain granular billing visibility on Amazon ECS, right-size workloads, and use auto scaling, AWS Graviton, and Spot Instances to optimize costs further.

In this session, we take a deeper look at Kubernetes HPA, VPA, and various custom metrics depending on the workload. We also perform a live demo of Kubernetes Cluster Autoscaler on AWS.