Define and enforce security architecture for AI platforms and agent-based solutions (e.g., Copilot Studio, Azure AI services), including identity and access controls for service principals and managed identities, least-privilege connector design, data protection and prompt handling safeguards, logging and traceability of agent actions, and integration with enterprise data governance controls (e.g., Microsoft Purview). Demonstrated experience designing and governing security architecture for hybrid (on-premises and cloud) and multi-cloud environments, including segmentation, secure connectivity (VPN, ExpressRoute, Direct Connect equivalents), DNS/routing, egress controls, and cloud governance models (landing zones, guardrails, subscription/account strategy).