Andrew and Joe Mora announce a new CrewAI course on building complex multi-agent workflows. You'll learn to organize agents into connected crews, run parallel tasks, pass outputs conditionally between them, and use live human feedback to improve performance.
I'm thrilled to be back with Joe Mora, the founder of TrueAI, who's teaching practical multi AI agents and advanced use cases. With TrueAI, you learn how to build and deploy AI agents that actually work and you see how such systems are in production right now and delivering real value right now for businesses.
Thanks Andrew. I'm super excited to be back here and I'm going to teach you so much about this course. It's going to be super interesting.
Multi AI Agent systems use multiple AI agents working together to achieve a common goal or to solve a complex task. In the previous multi AI agent systems course, you learned how to build multi agent systems using CrewAI. In this course you learned how to create more advanced agent based apps and automations that use external tools or do performance testing. You also learned to train your AI crew using human feedback and you also build truths that use different large language models for different types of tasks and learn how to deploy them.
Yes, on our previous course we covered organizing agents in simpler ways where results from one agent are passed to the next one. And in this one we will learn how to organize agents into more complex workflows, for example connecting multiple groups of agents, getting agents to work in a hybrid setting, and getting some tasks done in parallel. We're going to talk about reporting results to the agents above them. And you're going to learn how to create pipelines by building multiple crews together. Using a pipeline, one of the crews will finish its tasks and then, based on a conditional logic, pass the output into another crew to take another set of tasks.
After developing an application, you often want to measure so that you can continually improve its performance. In this course, you use testing and training to do this. Testing helps you create evaluations or evals for your crews by using another LLM to judge their performance and training in this context means providing live human feedback to your crews during operation so they can improve their behavior over time. For example, you may not be happy with the format of the report that is generated, or maybe your research agent is missing a link to the resources it has used. You can provide feedback on each task and have your agents learn from this feedback.
This course is going to be packed with exciting topics like automated project planning, lead scoring, support, data reporting, and content creation at scale. It's going to be a lot of fun and you will see how you can extend these use cases through your own work. I hope you join me on this one SA.
