Hands-On Practice
From the course AutoGen: Multi-Agent Conversations
Put what you have learned in this course into practice with the structured exercise below. Working through it end-to-end is the single best way to consolidate the concepts.
Exercise: Autogen Multi Agent in the Real World
Pick one realistic scenario from your own work, study, or curiosity and apply the techniques covered in earlier lessons. Aim for a small, completable artifact — a notebook, a working script, a written design, or a side-by-side comparison.
Suggested Workflow
- Frame the goal: write a one-sentence problem statement that the techniques in this course can address.
- Pick the right tool: revisit the lessons above and choose the specific method, model, or pattern that best fits.
- Build an MVP: produce the simplest version that proves the approach works. Keep it small.
- Evaluate: how did it perform? What surprised you? What broke?
- Iterate: refine one decision (prompt, parameter, tool) and re-evaluate.
Self-Check Questions
- Could you explain the chosen approach to a teammate in under two minutes?
- Where might this break in production? What guardrails would you add?
- What is the next thing you would try if you had another day?
Stretch Goal
Document your run as a short write-up. Compare against alternative approaches and note the trade-offs.
This practice loop converts passive reading into durable skill — the exact transfer-of-learning step most learners skip.
Concrete Starter Project: Build a Code Review Agent
Goal: Create a two-agent AutoGen system where an AssistantAgent acts as a senior engineer reviewing Python code submitted by the user.
Steps:
- Define an AssistantAgent with a system message specifying code review criteria (security, performance, readability, test coverage).
- Create a UserProxyAgent with
human_input_mode="TERMINATE"andmax_consecutive_auto_reply=3. - Submit a sample Python file with a few deliberate issues (missing error handling, hardcoded credentials, no docstrings).
- Observe how the agent identifies issues and suggests improvements.
- Respond to the agent's feedback and watch it refine its suggestions.
Extend it:
- Add a second reviewer agent specializing in security who flags any instance of hardcoded secrets or SQL injection risks.
- Add a
register_functiontool that runsflake8linting and returns the output to the agent. - Log all agent messages to a JSON file for later analysis.
What you will learn: How to configure agent roles, termination conditions, and tool registration to build a useful, production-safe multi-agent workflow.
Key Takeaways
- •Hands-on application converts course concepts into real skill
- •Start small with an MVP before optimising
- •Self-check questions catch gaps in understanding before they ship to production
Progress is kept in this browser only.