Your AI agents don't care about the policy PDF. They're acting right now — making calls, hitting APIs — and legacy IAM is oblivious. The challenge isn't writing rules on paper; it's engineering gates that actually stop bad behavior at runtime. AI Governance Platform Engineering (AIGPE) treats governance like a distributed-systems problem, not a paperwork exercise.
Lock in your spot for the first cohort. No fluff.
Govern where the agent acts, not where the policy is filed. Runtime gates, not shelfware.
Non-human identities (NHIs), reachable scope, and blast-radius — things traditional IAM was never built to express.
Tamper-evident ledgers and offline verification. Proof you can hand to an auditor without sweating through your shirt.
If a gate breaks, it needs to look broken. No fake green checks. No vaporware.
Three levels, each earned: Foundation → Practitioner → Architect. The ethos is baked in — readiness, not certification; confidence, not certainty; earned trust, not lock-in.
Literacy — the shared language and the mental model. Where governance attaches, and why runtime is the answer.
Build — engineer the deterministic gates, the identity binding, and the evidence ledger into a real agent platform.
Design & lead — set the autonomy boundaries, the readiness crosswalk, and the governance program across teams.
AIGPE is the people-side of the work Cognita ships: the governance and evidence layer for agentic systems. Read the field it's built on before you register.
Builds on and credits the Agentic Development Platform work (Weave Intelligence), the “path” concept (Charity Majors), and the runtime-governance argument (N. Kenney).