In this course, you'll get hands-on with AI agents and move quickly from theory to practice. You'll learn how to design agents that follow robust, precise instructions, acting purposefully and resisting manipulation. We explore applications in data and development environments, from code generation and execution to complex matching and reasoning tasks. You'll also learn how an agent can efficiently retrieve knowledge using RAG (Retrieval-Augmented Generation). Additionally, we cover the social, ethical, and legal aspects of working with agents. All implementations take place in Azure AI Foundry, a secure Microsoft-managed environment supporting multiple AI models, the same technology stack underlying tools like ChatGPT and Copilot.

What you'll learn

• The core principles of AI agents and their applications.
• How to write effective, safe, and reliable agent instructions.
• How agents retrieve knowledge using RAG models.
• Implementing agents in a Python development context.
• The social, ethical, and legal dimensions of AI agents.

Programme

Part 1 – Introduction to AI and ChatGPT

• What is AI, how does ChatGPT work, and what are its capabilities?

Part 2 – Introduction to Agents

• Concepts, capabilities, and limitations of AI agents.

Part 3 – Building and Testing Agents

• Hands-on building in Azure AI Foundry using Python.

Part 4 – Societal Context of Agents & RAG

• Ethical, social, and legal considerations.

Part 5 – Multi-Agents I

• Collaboration and role distribution between agents.

Part 6 – Multi-Agents II (open exercise)

• Designing and building your own multi-agent system.

Part 7 – Q&A & Wrap-up

• Summary, use cases, and next steps.

‍

• Developers and data engineers.
• Data scientists looking to integrate AI into their workflows.
• Python developers interested in ChatGPT, Copilot, and AI automation.

Prerequisites

• Basic knowledge of Python (with a local installation).
• Some experience with ChatGPT or other language models is recommended.