Build long running AI agents that pause resume and never lose context with ADK

How do you build AI agents that survive weeks of idle time, container restarts, and complex cross-team handoffs? Traditional stateless loops fail under context pollution, soaring token costs, and hallucinated progress.
This session dives deep into Google's Agent Development Kit (ADK) to build robust, long-running agentic workflows. We’ll break down a production-ready, three-pillar architecture:
* Durable State Machines: Enforcing explicit state schemas instead of relying on raw chat history
* Event-Driven Resumption: Waking dormant agents via FastAPI webhooks, allowing scale-to-zero during idle time
* Multi-Agent Coordination: Delegating tasks to specialized, focused sub-agents to keep reasoning sharp
Walk away with a practical, code-first playbook to ship enterprise background agents