Agentic Sandboxing & Infrastructure
These labs are about what Agents run on: isolating them in sandboxes (gVisor and microVMs) with Agent Substrate, and controlling what an Agent can reach, who can call it, and what it costs.
2 scenarios, 17 labs
This tutorial contains full, real-world scenarios for Agentic infrastructure & workloads. As the industry moves from standard workloads and microservices to "AI all the things", having a few key scenarios that engineers can use to not only learn about Agentic AI, but get ideas on what they can implement.

These labs are about what Agents run on: isolating them in sandboxes (gVisor and microVMs) with Agent Substrate, and controlling what an Agent can reach, who can call it, and what it costs.
2 scenarios, 17 labs
These labs are about building Agents and running them on Kubernetes: Models, MCP tools, an AI Gateway in front of the traffic, evals, and how much autonomy you hand over.
3 scenarios, 14 labs