Throughout the book, I focused on the theory of data engineering, its practical implementation on AWS, and where to apply generative and agentic AI to add significant value. I focused on the theory and fundamentals of building AI agents for data engineering. I saved the mechanics of building agentic systems for the last chapter of the book for exactly the following reason. The pace of innovation is so severe that just during the time it took me to write this book—admittedly longer than anticipated—models were augmented with agentic systems capable of tool use and sophisticated deep reasoning. Model context protocol (MCP) was open sourced by Anthropic. AWS released the Strands Agents SDK and its groundbreaking toolkit of agentic primitives, Bedrock AgentCore. While this volatility will continue, we can see the formation of a long-term agentic AI technology strategy driven by the industry and wider developer community. In this chapter, I dive deep into the mechanics of implementing agentic AI for data engineering on AWS. I demonstrate how to enhance existing APIs and data sources with MCP to enable agentic AI integration. I’ll build agents trained for data engineering tasks using the Strands Agents SDK. I’ll show you how easy it is to deploy AI agents to production with Bedrock AgentCore.