Mission: MCPossible — Deploying Real Agents with RAG and Tools at the Edge

From chatbots to agents that can actually get stuff done, LLMs are quickly growing up—but giving them tools and external knowledge they weren't trained on isn’t as straightforward as it seems. In this session, I’ll walk through how LLMs like Claude uses MCP (Model Context Protocol) to call external tools, how to automate document retrieval with AutoRAG, and how developers can combine both to build useful, reliable AI workflows. We’ll look at how to spin up your own tools with just a URL, make agents smarter with RAG, and avoid some common edge-case failures (pun intended). If you’re RAG-curious or already deep in agent land, this talk will help you ship something real.