Overview
The Enterprise AI Architect will play a vital role in leading the AI architecture workstream for a major transformation programme within a UK-based technology company. This position focuses on creating a buildable architecture framework that supports safe and effective AI implementation, guided by a structured three-zone model. The architect will collaborate closely with engineering teams to ensure the proposed architecture meets practical application requirements and is ready for internal handover.
Responsibilities
- Review the draft architecture and deliver a detailed multi-layer design, concluding with a build specification.
- Lead the evaluation and selection of AI gateway options, potentially including LiteLLM, Kong, and Azure APIM.
- Design the request pipeline, environments, egress controls, and Entra ID identity management, including agent identity.
- Define supply-chain controls and strategies for model provider, residency, and isolation.
- Establish a permission-aware grounding approach ensuring AI access is limited to entitled data.
- Develop data loss prevention (DLP) and guardrail policies, alongside security control specifications and threat models.
- Oversee governance processes from idea to production, including managing the project register, release gates, and approved tools.
- Deliver reference patterns for key initiatives such as AI-assisted software delivery and ML routing for customer service agents.
Requirements
- Essential experience in enterprise architecture and implementing Zero Trust principles.
- In-depth knowledge of Entra ID and conditional access policies.
- Strong understanding of egress and network controls in production environments.
- Experience with LLM gateways, model routing, and supply-chain controls.
- Desirable experience with running an LLM gateway in a production setting.
- Knowledge of regulated access governance over data and security frameworks.
- Experience in conducting enterprise security reviews and chairing design forums.