Ashok Srirama

AI agents are easy to prototype but hard to run in production. They have unique compute requirements, isolation between tenants, networking that connects to tools and APIs, and infrastructure that scales with unpredictable workloads.

If you run agents, multi-tenant services, or regulated workloads on shared infrastructure, the isolation boundary is an important security consideration. This session shows how to design that boundary using VM-level isolation on Amazon ECS.

Building secure and resilient networks for Amazon ECS with AWS Fargate has never been easier. In this chalk talk, learn about connecting microservices with AWS' native application networking tools, including Amazon ECS Service Connect and Amazon VPC Lattice.

Many organizations are reinventing their Kubernetes environments to efficiently deploy generative AI workloads, including distributed training and inference APIs for applications like text generation, image generation, or other use cases.

Building architecture based on best practices helps organizations build performant, cost-efficient, and secure applications. This chalk talk provides insights into Amazon ECS best practices grounded in the AWS Well-Architected Framework.