Building an Agent Runtime With Open Infrastructure
Talkat KubeCon + CloudNativeCon + OpenInfra Summit + PyTorch Conference China 2026 · 2026-09-07..09-09·
summary
Describes how to build a secure, short-lived agent runtime by combining existing open-source projects: Kubernetes Agent Sandbox for lifecycle management, Kata Containers for guest-kernel isolation, and upstream container/image delivery and execution tooling for efficient distribution of images and model artifacts. Explains how these components work together to create on-demand working environments, isolate untrusted code, and release resources when tasks finish.
AI-written summary · about
related talks
14mTaming Agentic AI: Run Rogue Code Safely on Kubernetes with Agent Sandbox
36mHow to Safely Deploy AI Agents that Write and Manage your IaC
29mSandbox for Agentic Application With Cloud Native Stack
28mAgent Runtime Security: Enforce, Attest, Decide
25mAgent Based Kubernetes Installation
39mTrust, but Sandbox: Securing AI Agents in Kubernetes
23mAgents in Kubernetes 101: A Crash Course on Cloud Native AI Agents
29mKubernetes, created for fast deployment, connectivity and ease of use
