Skip to Content
View all events

MCP Bootcamp: Building AI Agents with Model Context Protocol

Published by O'Reilly Media, Inc.

Intermediate content levelIntermediate

Connecting LLMs to real-world data through standardized interfaces

Course outcomes

  • Create simple agents that can leverage tools powered by MCP
  • Build fully functional MCP servers that connect LLMs to external data sources and tools
  • Design and implement AI workflows that use MCP to provide context-aware responses
  • Create agents with Claude Agents SDK that can use MCP tools to perform tasks

Course description

The Model Context Protocol (MCP) is an open protocol that standardizes how AI applications connect to data sources and tools. Think of it as a USB-C port for AI—a universal connector that lets large language models seamlessly integrate with external systems.

Join expert Lucas Soares to learn how to build MCP servers that expose tools, resources, and prompts to LLMs like Claude. You’ll explore how to create AI workflows that leverage external context and tools through MCP, enabling more powerful and context-aware AI assistants. Through practical examples, you’ll build agents that can leverage MCP servers for file systems and have the ability to leverage APIs using the Claude Agents SDK. By the end of the course, you’ll have the skills to build sophisticated AI-augmented workflows that connect LLMs to the data and tools they need to be truly useful.

What you’ll learn and how you can apply it

  • Design and implement composable agent architectures using MCP
  • Understand the MCP architecture and how it enables AI applications to access external context
  • Build MCP servers that expose tools, resources, and prompts to LLMs

This live event is for you because...

  • You’re a software engineer, AI engineer, data scientist, or technical product manager.
  • You work with large language models and AI-powered applications.
  • You want to build AI workflows that connect LLMs to real-world data and tools.

Prerequisites

  • A computer with Python 3.10+ and the MCP SDK installed (optionally, Node.js for JavaScript examples)
  • Access to Claude API (free tier sufficient) for testing MCP integrations, or access to Claude Desktop
  • Basic Python or JavaScript programming experience
  • Familiarity with REST APIs and web services
  • Understanding of basic LLM concepts
  • No prior experience with MCP required

Recommended preparation:

Recommended follow-up:

Schedule

The time frames are only estimates and may vary according to how the class is progressing.

Day 1

Introduction to Agents & MCP (40 minutes)

  • Presentation: Agent Basics
  • Hands-on: Introduction to agents from scratch
  • Presentation: Where does MCP fit into the Agentic picture?
  • Demo: Examples of Agents with MCP (Claude Code)
  • Introduction to MCP for Agent Builders (40 minutes)
  • Presentation: MCP architecture & core concepts—tools, resources, prompts; MCP clients vs. servers; the MCP ecosystem
  • Presentation: MCP concepts for building Agents—clients vs. servers; capabilities (tools, resources, prompts); schemas & validation; local dev workflow

Q&A (10 minutes)

Break (10 minutes)

MCP Quickstart: Build Your First Server (50 minutes)

  • Presentation: Dev environment setup; MCP SDK project skeleton; define a minimal tool & prompt; run & inspect requests/responses
  • Hands-on exercise: Create a simple "Utilities" MCP server and a starter prompt
  • Demo: Use your MCP server with pre-built tools (Claude Code, Cursor)

Q&A (10 minutes)

Break (10 minutes)

Setup & First Working Agent with Claude Agent SDK (60 minutes)

  • Presentation: Intro to Claude Agent SDK + MCP
  • Hands-on exercise: Build a minimal “File Reader” agent—pip install claude-agent-sdk; start a streaming session; attach an MCP Filesystem server; list tools; invoke read + return a summary.

Q&A (10 minutes)

Day 2

Warm-up / Day-1 Recap (15 minutes)

  • Building an Automation Agent with Claude Agent SDK + MCP (60 minutes)
  • Presentation: Automating automations with Claude Agents SDK + MCP
  • Hands-on: Implement an agent that creates and tests automation scripts automatically from database

Q&A (10 minutes)

Break (10 minutes)

Claude Code + MCP as an Agentic Generalist System (60 minutes)

  • Presentation: Agentic Productivity with Claude Code + MCP (Framework & Workflow)
  • Hands-on: How to setup Claude Code as a Generalist System Leveraging Core MCP Servers

Q&A (10 minutes)

Break (5 minutes)

MCP Hacks & Tips for Devs (50 minutes)

  • Presentation: Practical patterns—multi-server layouts, schema validation in practice, logging & observability, local dev loops, safety guards, perf tips

Demo: MCP for Everything with Claude Code (10 minutes)

  • Build a centralized dashboard hub (Claude Code + MCP): attach servers, wire prompts/tools/resources, quick operator shortcuts

Final Recap & Q&A (10 minutes)

Your Instructor

Lucas Soares

Lucas Soares is a machine learning engineer who has worked at K1 Digital and Biometrid, where he developed computer vision and NLP models for applications such as document verification, OCR-based applications, and recommender systems. Lucas has also developed various ML models, including neural networks, Siamese networks, convolutional neural networks, LSTMs, and genetic algorithms.

linkedinXlinksearch

Skills covered

  • Claude
  • MLOps