AutoGen's Successor Microsoft Agent Framework, 17 Exercises

What you'll learn

---

Understand what happened to AutoGen and what MAF inherited from it, and explain the three sticking points when migrating AutoGen code to MAF.
Understand ChatClient / Agent / Session, shape a personality with instructions, keep conversations going with Session, and separate history per user.
Build custom tools with `@tool`, understand how type hints and docstrings double as the AI's instruction manual, and build an agent that chooses between tools.
Explain the five coordination patterns - Sequential, Concurrent, Handoff, GroupChat, and Magentic - with diagrams, and decide which one fits your own problem.
Cut processing time with parallel execution, and implement a Handoff that routes requests to the right specialist - adding a new desk costs almost nothing.
Build a directed graph with WorkflowBuilder, design branching with switch-case and fan-out / fan-in, and catch bugs before execution via type checking.
Pause a Workflow for human approval with `request_info`, and resume a crashed run with `FileCheckpointStorage` - features AutoGen never had.
Merge classification, specialist routing, an approval gate, and storage into one pipeline that holds up to audit and approval Requirements

---

.
When an official builder fails with a provider, learn to tell from the error message whether the framework or API is at fault, and build a workaround.
Requirements

---

Basic Python syntax - variables, functions, classes, enough to read imports and decorators. No async experience needed; Section 0 covers MAF's conventions.
A PC with internet and a modern browser (Chrome recommended) is all you need. The course uses CodeSandbox, so no Python install or setup is required.
Ability to get one Gemini API key from Google AI Studio (free tier, no credit card needed - Section 0 walks through the steps).
No Azure subscription required. Although this is a Microsoft framework, the course is designed so the free Gemini tier alone gets you through.
No prior AutoGen or Semantic Kernel experience needed. A migration guide for experienced users is in the bonus section; first-timers can start right in.
Description

---

This course contains the use of artificial intelligence.
"I set out to learn AutoGen, only to discover it had entered maintenance mode" - if that's you, this course was made exactly for that situation.
AutoGen moved into maintenance mode in October 2025, and its successor, Microsoft Agent Framework (MAF), reached general availability in April 2026. Microsoft officially describes MAF as the direct successor to both AutoGen and Semantic Kernel, built by the same team. It combines AutoGen's simple agent abstraction with Semantic Kernel's enterprise features - session management, type safety, and checkpointing. This course covers that successor framework, and includes a migration section for developers coming from AutoGen.
MAF rests on two pillars: Agents (autonomous) and Workflows (controlled). One lets the AI make the call; the other lets a human fix the execution order. Knowing when to use which is exactly what turns a demo into a production system. This course builds both up through 17 hands-on exercises.
The course follows an 8-part roadmap. Section 1 covers the three-layer structure of ChatClient, Agent, and Session; Section 2 teaches how to give an agent tools with `@tool`. Section 3 is the first major milestone: multi-agent orchestration. It diagrams the five standard orchestration patterns, then has you build parallel review and a reception-to-specialist handoff yourself.
Section 4 opens the second major milestone: Workflows. You move from writing procedurally with `@workflow` and `@step`, to building a directed graph with `WorkflowBuilder`, through conditional branching (switch-case) and parallelism (fan-out / fan-in). A type mismatch fails before execution even starts - you'll experience catching bugs without calling an LLM even once. Section 5 covers three features every real deployment needs: human-in-the-loop approval via `request_info`, checkpointing that resumes a failed run partway through, and visualization with DevUI.
The final chapter, Section 6, brings everything together. You classify an inquiry, have the right department draft a reply, and save it as Markdown once a human approves it - completing a pipeline built to withstand audit and approval Requirements

---

.
Every line of code in this course has been verified against the live API. Along the way, cases turned up where the official builder simply does not work with Gemini. This course doesn't hide that - you'll diagnose the error message, work out why, and rebuild it yourself. Learning to handle a broken official example on your own is not something most introductory courses teach.
No environment setup required - you complete the entire course in a browser with CodeSandbox. And you don't need an Azure subscription: the free Gemini tier alone gets you all the way through, with no credit card required. That makes this an unusually accessible way to learn MAF.
Who this course is for

---

Anyone who set out to learn AutoGen and was thrown to find it's in maintenance mode, wanting to know what to relearn and if prior knowledge still counts.
Developers actively writing AutoGen or Semantic Kernel code who are considering a move to Microsoft Agent Framework.
Engineers who can build a single AI agent but aren't sure how to design coordination between multiple agents or fold them into a business workflow.
Anyone who wants an AI agent in an internal system but is stuck on whether it's safe to hand it everything, and wants to design approval and audit trails.
Anyone who has learned LangChain or CrewAI and wants to see how Microsoft's official stack solves the same problems, to broaden their framework options.
Independent developers and students who want to try multi-agent development on the free tier alone, without paying for Azure.
Homepage

---