Build Your Own Team of Managed Agents - Jack Wotherspoon, Google DeepMind

Jack Wotherspoon, developer experience engineer at Google DeepMind, introduces Gemini Managed Agents: coding agents that run in cloud sandboxes from a single API call instead of on your laptop. He covers the Interactions API, the Antigravity harness, environments that several agents can share, uploaded skills and AGENTS.md files, a proxy that keeps secrets out of the sandbox, scheduled triggers and multi-agent workflows. Demos include an SEC filing researcher handing its work to a quant agent, and a stock research app built in AI Studio. Chapters 00:00 Intro: what are managed agents? 00:54 Why run agents in the cloud 02:13 Doing it yourself vs one API call 03:26 The Interactions API 04:32 How it works under the hood 05:22 Building blocks: harness, environment, skills, credentials, triggers 07:02 Choosing models, including Gemini 4 Argon 08:51 Interactions and environments 10:46 Tools, MCP servers and file mounts 12:21 Keeping secrets out of the sandbox 13:12 Triggers and scheduled agents 13:37 Multi-agent workflows 14:32 Demo: SEC research to quant analysis to report 16:34 Demo: managed agents inside an app 17:44 Getting started in AI Studio Jack Wotherspoon, Google DeepMind Join us in SF for the second annual AIE CODE summit: https://aie.codes our last flagship summit of 2026. Applications live now!