Agentic AI on AWS

This course teaches engineers how to design, reason about, and implement production-shaped agentic AI systems on Amazon Web Services (AWS). Participants use Amazon Bedrock for managed model access, open orchestration patterns for agent execution, and Model Context Protocol (MCP) as a standardized interface between agents and external tools and services.

The course establishes the architectural foundations of agentic systems, including control loops, tool use, context, memory, orchestration, reliability, evaluation, security, and governance. These concepts are then mapped to AWS services and current AWS agent infrastructure without making any single framework the architectural center of the course.

Participants learn how Amazon Bedrock AgentCore capabilities can support production agent runtimes, MCP-based tool integration, identity, observability, and operational controls. AgentCore Gateway is used to demonstrate how APIs, Lambda functions, AWS services, and existing MCP servers can be exposed through standardized MCP interfaces while maintaining AWS authentication and authorization boundaries.

By the end of the course, participants will be able to design and implement agentic workflows that are operable, secure, observable, cost-aware, resilient, and aligned with AWS production practices.