Building Production-Ready Custom AI Agents on AWS (Part 1/2)

This session focuses on the transition from AI proofs of concept to robust, scalable, and secure production deployments. Attendees will explore the core components of AI agents, including agent loops, state management, and prompt design using the Strands framework. The session also covers AWS Bedrock AgentCore, detailing production essentials such as runtime, code interpretation, memory, gateway, identity, and observability. A real-world demonstration illustrates how these components integrate to create functional and maintainable agents, providing practical insights for developers and architects.