From the course: Azure AI for Developers: Building AI Agents

Unlock this course with a free trial

Join today to access over 24,900 courses taught by industry experts.

AutoGen Swarm

AutoGen Swarm

- [Presenter] Swarm implements a team in which agents can hand off tasks to other agents based on their capabilities. It is a multi-agent design pattern first introduced by OpenAI in an experimental project. The key idea is to let agents delegate tasks to other agents using a special tool call, while all agents share the same message context. This enables agents to make local decisions about task planning, rather than relying on a central orchestrator such as in SelectorGroupChat. The speaker agent is selected based on the most recent HandoffMessage in the context. This requires each agent in the team to be able to generate HandoffMessages to signal which other agents that it hands off to. In this notebook, we again use the same multi-agent system that we have created in the previous video. We define our dalle agent, but this time we specify that it hands off to the planning agent through the hands of parameter and the system message. We then define our vision agent using the same…

Contents