Swarm Theory -----Multi-agent coordination, gone deep. This is a Agent Hangout Session.

​One agent is a tool. A hundred agents coordinating in real time? That's a system — and systems have failure modes, emergent behaviors, and design decisions that will make or break everything you're building.

​Swarm Theory is a technical deep-dive for builders, engineers, and architects who are done talking about agents in theory and ready to get into the architecture. We're pulling apart how multi-agent systems are actually designed — how tasks get routed, how context gets passed, how you prevent one broken agent from cascading into a full system failure.

​This is a discussion-first format. Bring your stack, your failures, and your half-baked orchestration ideas. The best conversations happen when people get specific.

​We'll get into: — Orchestration patterns — supervisor, peer-to-peer, hierarchical, and when each breaks down — Memory, context, and state — what gets shared across agents and what stays siloed — Tool routing, handoffs, and how to debug a swarm that's misbehaving — Real-world implementations — what's working in production right now

​Whether you're building with LangGraph, CrewAI, custom orchestration layers, or stitching it together yourself — if you're thinking in agents, this is your room.

​Drinks. Builders. Deep conversation. No slides required — just show up ready to think out loud. Hosted by PDX Hacks — Portland's weekly AI builder series.