AI Infrastructure Engineer

Key Responsibilities

- Platform Administration: Administer and configure the corporate AI platform estate at tenant level, including service enablement, environment strategy, admin role delegation, and controlled adoption of preview and general availability releases across Microsoft and non-Microsoft AI platforms.

- Agent Lifecycle Management: Operate the agent registry end to end - discovery and onboarding of agents (including shadow agents), ownership assignment, approval and publication workflows, certification, change control, and retirement - across Copilot Studio, Agent Builder, Power Platform, and third-party agent platforms.

- Identity and Access Requirements: Define the identity and access requirements of each agent and AI platform - intended scope, permissions, consent model, and least-privilege boundaries - and hand them over to the Entra identity team for implementation, then validate that the delivered configuration matches what was requested.

- Guardrails and Policy Enforcement: Translate approved corporate AI safeguards into enforceable technical controls and evidence their operation - data loss prevention, connector and environment boundaries, tenant isolation, admin consent, sensitivity labelling, and audit logging - in partnership with Security, Privacy, and Compliance.

- Automation: Automate provisioning, governance, remediation, and reporting using PowerShell 7, Microsoft Graph, and platform administration APIs and CLIs, with version-controlled, peer-reviewed, and documented code.

- Service Operations: Act as escalation point for AI platform incidents and service requests; define runbooks, monitoring, and service levels; and drive problem management through to root cause and permanent fix.

- Collaboration and Enablement: Partner with the Entra identity team, the AI Centre of Excellence, Security, Legal and Privacy, Data, and business units to onboard new agent use cases safely, and coach makers and agent owners on governance requirements and platform boundaries.

- Cost and Licence Management: Manage licensing, capacity, and consumption across the AI estate; deliver segregated cost allocation and chargeback reporting per platform and business unit; and identify optimisation and rationalisation opportunities.

- Usage Analytics and Adoption Reporting: Build and maintain dashboards covering agent inventory, usage, value, and risk signals, and report adoption, service health, and return on investment to platform owners and business stakeholders.

- Release Monitoring and Service Incorporation: Review the continuous flow of AI feature announcements across the platforms in scope - the Microsoft 365 Message center and roadmap, Power Platform, Copilot and Fabric release notes, and the updates published by OpenAI and Anthropic - assess the impact of each relevant change on configuration, governance controls, licensing, cost, and users, and turn that assessment into a documented decision to adopt, contain, or defer, followed by the pilots, policy updates, and communications that the decision requires.

- Documentation and Standards: Maintain governance standards, administration runbooks, architecture decision records, and maker-facing guidance that keep the platform supportable, transferable, and audit-ready.