Track · Foundry Infra
6 weeks · 4 missions · 0 in cohort
AI Agentic OS Infrastructure
Build the substrate other teams' agents run on.
§01 / The brief
Why this track exists.
The infrastructure track for the agentic era. You will design and operate the runtime, memory, eval and policy layer that makes agents safe to ship — not just fun to demo.
Runtime, memory, evals and policy for shipping agents in production.
§02 / Who it's for
You are…
- →Platform engineers
- →Senior backend devs
- →Infra-curious founders
§03 / What you'll learn
By the end…
- ✓Agent runtimes and orchestration
- ✓Tool & memory design
- ✓Evals that survive contact with prod
- ✓Policy, audit, and rollback
§05 / Missions
The work, in order.
- 01Runtime design
Pick the substrate. Define orchestration boundaries.
lesson - 02Tools & memory
Design tool surfaces and durable memory.
challenge - 03Evals in prod
Build evals that survive contact with reality.
challenge - 04Policy & rollback
Audit, policy and a believable rollback story.
peer_review
§06 / Tools
Your stack.
TypeScriptPythonPostgresTemporalOpenAIAnthropicLovable
Tools shift fast. We teach the workflow so you can swap any of these without losing a beat.
§07 / What you ship
Portfolio you can defend.
- ◆A multi-agent runtime running real tasks
- ◆An eval suite
- ◆An operational doc you'd hand to a successor
§08 / FAQ
Common questions.
How deep does this go?+
Deep — this is for people who'd rather build the platform than use it.
§09 / Related
Stack it with…
Next cohort