AI Security Architect (Enterprise LLM / Claude Code)

Project overview

Contract role in London to design and own end-to-end security architecture for Claude Code and future enterprise AI deployments. The architect will lead threat modelling and security architecture for agentic AI/LLM workloads, define enterprise guardrails and standards for AI tool usage, evaluate and select security tooling (e.g., DLP, API gateway, audit logging), and engage AI vendors (including Anthropic) on security controls and contractual requirements.

Requires strong expertise in Zero Trust/SASE and API security, STRIDE threat modelling, security architecture frameworks (SABSA/TOGAF), and knowledge of AI-specific risks (prompt injection, model inversion, data poisoning) plus regulatory awareness (GDPR, EU AI Act, NIST AI RMF).

Responsibilities

• Lead threat modelling and architecture design for Claude Code and agentic AI workloads.
• Define security standards, patterns, and guardrails for AI tool usage enterprise-wide.
• Evaluate and select security tooling (DLP, API gateway, audit logging) for AI use cases.
• Engage with Anthropic and other AI vendors on security controls and contractual requirements.
• Provide expert input to the AI Governance Committee.
• Define and evolve the security target operating model for AI.
• 8+ years in cybersecurity, with 2+ years in AI/ML security or LLM deployment.
• Deep knowledge of Zero Trust, SASE, and API security architectures.
• Experience with STRIDE threat modelling and security architecture frameworks (SABSA, TOGAF).
• Knowledge of relevant regulations: GDPR, EU AI Act, NIST AI RMF.
• Strong stakeholder communication; ability to present to C-suite.
• Experience with Anthropic Claude API or similar LLM platforms.

Requirements

• Security architecture (end-to-end design/ownership)
• Threat modelling (STRIDE)
• Zero Trust architecture
• SASE concepts/architecture
• API security architecture (authn/authz, gateways, rate limiting, abuse prevention)
• AI/LLM security for enterprise deployments (agentic AI workloads)
• AI-specific risk knowledge (prompt injection, model inversion, data poisoning)
• Security architecture frameworks (SABSA or TOGAF)
• Security standards/patterns/guardrails definition
• Security tooling evaluation/selection for AI use cases (DLP, API gateway, audit logging)
• Regulatory/security governance knowledge (GDPR, EU AI Act, NIST AI RMF)
• Stakeholder communication and C-suite presentation
• Cybersecurity experience (8+ years) with AI/ML security or LLM deployment exposure (2+ years)

Project details

• Start: flexible
• Remote: remote
• Location: United Kingdom, GB