Platform teams are under pressure as AI tools like Copilot accelerate code generation, creating governance and scalability challenges. AI Platform Engineering covers two mandates: augmenting IDPs with AI capabilities, and building governed infrastructure for AI/ML workloads. A modern stack layers Kubernetes (substrate), Backstage (portal), and OpenChoreo (control plane). OpenChoreo is an open-source IDP that declaratively manages Projects, Components, Endpoints, and Connections — automatically generating network policies via Cilium, CI pipelines via Argo Workflows, observability via Prometheus/Fluent Bit, and Backstage catalog entries. It also ships a native MCP server enabling AI tools like Copilot, Cursor, or Claude to manage platform resources through natural language. The article contrasts raw Kubernetes deployment (days of manual YAML work) against OpenChoreo's declarative path (one Project, one Component, one Endpoint). Key recommendations include structural policy enforcement, GitOps audit trails, dedicated observability planes, and self-governing service catalogs to safely absorb AI-generated code at scale.