Skip to Content
View all events

Build Production-Ready AI Agents

Published by Pearson

Intermediate content levelIntermediate

Use LangGraph, CrewAI, and Claude MCP to build autonomous agents that both think and act

  • Build real autonomous agents with LangGraph's stateful workflows and CrewAI's multi-agent orchestration.
  • Connect agents to everything using Claude MCP servers for databases, APIs, and tools.
  • Deploy production-ready systems that go beyond demos—actual agents doing actual work.

Production-ready AI agents go beyond chatbots--they plan, execute, maintain state, and make decisions independently. This course teaches you to build autonomous agents using the enterprise stack that's actually shipping: LangGraph for stateful orchestration, CrewAI for multi-agent collaboration, and Claude's Model Context Protocol (MCP) for universal tool connectivity.

You'll build a working software development team of AI agents--Product Manager, Developer, and QA Tester--that collaborate to complete real tasks. No toy demos or "hello world" examples. By session's end, you'll understand when agents beat traditional automation, how to architect reliable agent systems, and most importantly, how to avoid the pitfalls that tank 90% of agent projects. This is for developers who need to deliver production systems, not experiments.

What you’ll learn and how you can apply it

By the end of this live online course, you'll be able to:

  • Design and deploy stateful agents using LangGraph's workflow engine for complex, multi-step reasoning.
  • Orchestrate multi-agent systems with CrewAI for collaborative problem-solving.
  • Build MCP servers to connect agents to any database, API, or tool in your stack.
  • Evaluate agent ROI with a framework for when to use agents vs. traditional automation vs. humans.

This live event is for you because...

  • You're a developer who's heard the "agents" hype and needs to separate reality from marketing.
  • You're a software architect who has experimented with LangChain or similar tools but struggled to move from demos to production.
  • You’re a technical manager who needs to make informed decisions about agent adoption for your team or organization.

Prerequisites

  • Python proficiency including async/await patterns and basic API development.
  • LLM fundamentals--you've used ChatGPT, Claude, or similar and understand prompting basics.
  • Command-line comfort with pip, virtual environments, and environment variables.
  • API experience with REST endpoints and JSON—we'll be connecting lots of services.

Course Setup

  • Python 3.11+ with pip and venv configured
  • Claude API key (Pro account recommended)
  • OpenAI API key (for comparison demos)
  • VS Code or preferred Python IDE
  • Docker Desktop (for MCP server deployment)
  • Course repo: github.com/timothywarner-org

Recommended Preparation

Recommended Follow-up

Schedule

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

Segment 1: Agents vs "Agents" - Foundations & Architecture (50 minutes)

  • What makes a true autonomous agent vs. a prompted chatbot
  • LangGraph fundamentals: nodes, edges, and state management
  • Demo: Building your first stateful agent with memory and decision-making
  • CrewAI introduction: role-based agents and delegation patterns
  • Cost reality check: When agents are worth it (and when they're not)
  • Mini-exercise: Design an agent workflow for your use case

Q&A (10 minutes)

Break (5 minutes)

Segment 2: LangGraph Deep Dive - Stateful Workflows (50 minutes)

  • Building complex graphs with conditional edges and loops
  • State persistence and checkpointing for long-running agents
  • Demo: Code review agent that maintains context across files
  • Error handling and fallback strategies when agents fail
  • Human-in-the-loop patterns for critical decisions
  • Mini-exercise: Add state management to a basic agent

Q&A (10 minutes)

Break (5 minutes)

Segment 3: Multi-Agent Orchestration with CrewAI (50 minutes)

  • Creating specialized agents: PM, Developer, QA Tester roles
  • Agent communication and delegation strategies
  • Demo: Software team completing a feature request end-to-end
  • Managing agent conflicts and contradictions
  • Performance optimization: parallel vs. sequential execution
  • Mini-exercise: Build a two-agent collaboration

Q&A (10 minutes)

Break (5 minutes)

Segment 4: MCP Servers & Production Deployment (50 minutes)

  • Claude MCP architecture: universal tool protocol
  • Demo: Building MCP server for database access
  • Connecting agents to GitHub, Slack, and internal APIs
  • Production considerations: rate limits, costs, monitoring
  • Agent evaluation framework: metrics that matter
  • When NOT to use agents: antipatterns and alternatives
  • Wrap-up exercise: Architecture review of complete system
  • Q&A (5 minutes)

Overall wrap-up and next steps (5 minutes)

Your Instructor

  • Tim Warner

    Tim Warner has been a Microsoft MVP in Azure AI and Cloud/Datacenter Management for 6 years and a Microsoft Certified Trainer for more than 25 years. His O'Reilly Live Training classes on generative AI, GitHub, DevOps, data engineering, cloud computing, and Microsoft certification reach hundreds of thousands of students around the world. He's written for Microsoft Press, presented at Microsoft Ignite, and contributed to several Microsoft open-source projects. You can connect with Tim on LinkedIn: timw.info/li.

Skill covered

Generative AI