Multi-Agent Systems: A Practitioner's Workshop

Most "multi-agent" demos collapse the moment they meet a real workload. This workshop is the antidote: 3 hours of structured, hands-on design with the patterns that actually hold up. You'll work through: supervisor vs swarm topologies, when to use a router agent, how to design tool boundaries, fault recovery for long-running agent loops, and the eval rigs you need to catch silent failures. Bring a laptop. By the end, you'll have a working multi-agent prototype shipping back to your IDE.