Master Agentic AI with Python in This 4-Hour Video Tutorial

Introduction

Agentic AI, often described as AI with autonomy and decision-making capabilities, is revolutionizing how businesses leverage artificial intelligence. A new four-hour video workshop by Jon Krohn and Edward Donner demystifies this technology, providing a hands-on guide to building agentic AI systems in Python.

Key Topics Covered

• Defining Agents: AI agents use LLM outputs to control workflows, enabling autonomy beyond predefined tasks.
• Business Value: The workshop highlights the potential of agentic AI to transform industries by 2025, with LLMs achieving dramatic improvements in benchmarks like Humanity's Last Exam (HLE).
• Frameworks Explored:
  • Model Context Protocol (MCP): A universal connector for agentic applications.
  • OpenAI Agents SDK: Lightweight and flexible for research.
  • CrewAI: Designed for multi-agent systems.
  • LangGraph and Microsoft Autogen for advanced use cases.

Hands-On Projects

• Deep Research Replication: Using OpenAI's SDK to perform web searches and generate reports.
• Autonomous Software Engineering: Building a team of AI agents to write, test, and even generate UIs with CrewAI.
• Simulated Trading Agents: Leveraging MCP to access real-time data and make trading decisions.

Takeaways

By the end of the workshop, viewers will:

• Understand core agentic AI concepts and frameworks.
• Implement multi-agent systems with tools like CrewAI and MCP.
• Mitigate risks like unpredictability and high costs.

For those eager to dive into agentic AI, the full video and GitHub code are available for further exploration.

By Matthew Mayo, Managing Editor at KDnuggets.