Build long-running agents with Google’s Agentic Stack | The Agent Factory
Google adk 2.0 Agent studio Long Running agents Agentic ai Google cloud ai Multi Agent systems Agent persistence Event Driven architecture Addy osmani Ai agent architecture
Google Cloud Tech demonstrates building production-grade long-running agents using ADK 2.0 and Agent Studio. The video covers three essential patterns: durable state checkpoints for persistence, event-driven dormancy for true sleep, and separated evaluation for quality control. Viewers see live demos of an HR onboarding coordinator, autonomous OS builder, and a 3D Blender-generated video store—all running over multi-day workflows.
Related videos
Your AI agent is forgetful. Here’s how to give it a brain.
The future of Cloud AI: Mastering MCP servers, Gemini, and agentic workflows
How we built 1,000 AI agents that run a marathon
Introducing Work IQ in the IQ Series
![[Demo] Autonomous ML Reliability - Data Center Network](https://i1.ytimg.com/vi/0yKGILWlngY/hqdefault.jpg)
[Demo] Autonomous ML Reliability - Data Center Network
CNCF On-Demand: Autonomous Agents on K8s – Durable Execution for AI
Tuning custom Gemma models for high speed computer vision
The ABCs of agent building
Mastering AI stacks for software engineers
Building distributed multi-agent systems
Agent-first workflows from prompt to production
