This session is aimed at beginner to intermediate users. Participants will explore a prototype multi-agent workflow and learn about the architectures, frameworks, and techniques used to develop, trace, debug, and monitor multi-agent systems. The session will also cover the importance of model selection and model capabilities, as well as methods for monitoring token usage and workflow execution. A prototype demonstration will showcase the behaviour of an orchestrator agent coordinating a set of specialised agents within a structured workflow to create complex shapes using the Turtlesim 2D simulator. 

Attendance at Session 1 is highly encouraged. A basic understanding of agentic systems will also be beneficial. 

Resources and Tools

Github Link - Session 2