The Mastering Multi-Agent Development with AutoGen course is a well-structured and comprehensive program for AI practitioners, developers, and AI automation enthusiasts. It effectively balances theoretical knowledge with hands-on implementation, making it suitable for both beginners and experienced professionals. The focus on structured conversations, cost efficiency, security, and advanced AI automation makes this a must-attend course for those looking to master multi-agent systems with AutoGen. Highly recommended for anyone interested in multi-agent AI development and automation.

The course introduces Agents, and AutoGen concepts clearly. The instructor walks us through the concepts with hands-on material/use-cases that cover various conversation patterns. Overall, pretty happy with the course as an introduction to the AutoGen Agentic Framework.

The course is detailed and high quality, but unfortunately, it's outdated due to the latest version of Autogen. However, it's still useful for building a solid foundational understanding.

Having taken several other courses on Autogen, I found this course valuable although confusing on occasion. Especially when getting into the complex patterns (groupchat, nested chat), the presenter seemed to wing it while explaining the output, glossing over inconsistencies and not catching errors in the output. Also the method of initiating chats in code was inconsistent: in a group chat, why would you initiate on an assistant agent rather than on the group manager? Also i found some of the diagrams less than helpful. Lastlly, some exam questions were just strange, quizzing on example code aspects instead of general concepts and understanding. Overall, a mixed bag.

Mostly worthless. There's very little explanation and lot of just browsing incomplete code (videos are so zoomed in you literally can't see the complete code). The instructor doesn't give deep explanations most of the time, leaving it at "now we'll use class X instead of Y" with 0 reasons given for why. The examples are extremely bad. There's no motivation given for most of them as to why multiple agents are any better than a single agent (likely there's no advantage for the examples chosen). Some of the examples are also broken, with the prompts being so vague that the models don't know who's the user and who's the assistant - and them both choosing to be the assistant.