Senior Director, AI ML and Cloud Security

Description

Senior Director role leading AI, ML and cloud security at a global biopharma company, owning the security function that designs, builds and governs AI and ML security capabilities.

The function covers an agentic framework modernising security services across the Chief Security Office, the AI and ML platform, and a data fabric that the Cyber Architecture and Solutions team designs and builds and a managed service vendor operates.

You will own AI security standards that bind all company AI initiatives, not only those in Cyber Architecture and Solutions, and cloud security governance across the Azure and GCP estate.

The function has 17 roles across the UK, India and the US, covering AI architecture, AI engineering, AI platform and cloud security, business-facing AI advisory, and enterprise AI security authority.

The team is transforming security itself by building a multi-agent platform that moves Chief Security Office services from rule-based to AI-driven operation. It owns the agent layer, LLM orchestration, the AI and ML platform and the data fabric pipeline, and it acts as the link between business units, R&D and the technical AI function.

This is not an AI governance role with an engineering team attached. You will set agent authorisation models and trust boundaries where no industry precedent exists, and decide whether an LLM-based control is reliable enough for a security outcome in a GxP environment where failures can reach medicine supply and clinical data.

Engineers work in LangChain, LangGraph, vector databases and production Kubernetes and will expect you to follow the technical detail and challenge it. Candidates whose AI experience is limited to policy frameworks, maturity models and vendor assessments are unlikely to fit, and this will be tested early in the interview. Candidates with deep AI engineering but no experience holding a security position against a funded programme under delivery pressure will find the role difficult.

Responsibilities

• Set the three-year plan to move Chief Security Office services from rule-based to agent-operated, and defend the investment case to the CISO and the company Tech leadership.
• Own the architecture, engineering and secure delivery of the agentic framework across the Chief Security Office, including agent authorisation models, tool-invocation control, human-in-the-loop gates, trust boundaries and runtime monitoring, built in from the start.
• Own the pipeline from design to build to handover: specifications engineered faithfully, the data fabric built and tested to specification, and vendor handover packages strong enough for a managed service to run safely from day one.
• Own the AI security standards, reference architectures and threat models that apply across Cyber Architecture and Solutions and every company AI initiative. No AI capability is architected, built or released without meeting them.
• Lead a 17-role function across five sub-functions and three countries, including team leaders, and own its capability and succession planning in a scarce skills market.
• Own AI platform security governance, including agent risk tiering, LLM and API exposure standards, AI-specific incident response, and the shadow AI discovery programme across the business.
• Own cloud security governance across Azure and GCP on an advisory and standards-setting basis, using standards and escalation to exercise authority while platform teams keep remediation.
• Own adversarial AI defence, including AI red teaming, prompt and model attack testing, and the mandatory remediation process that follows, informed by MITRE ATLAS and current research.
• Turn the EU AI Act, NIST AI RMF and ISO/IEC 42001 into architectural requirements ahead of mandated deadlines, not in response to audit findings.
• Represent the function to the CISO, the company Tech leadership, the Enterprise AI programme, auditors and regulators, and own board-level escalation of material AI security risk.
• Own the commercial relationship with the managed service vendor that runs the agentic framework and data fabric, including the service definition, handover gates and exit plan.

Requirements

• Significant senior leadership experience in cyber security or AI engineering, including senior leadership in a regulated environment, with a record of setting AI security strategy and roadmaps with measurable outcomes across AI architecture, AI engineering, AI platform and cloud security, and business-facing AI.
• Experience designing and delivering an agentic AI framework into production at enterprise scale, covering multi-agent orchestration, tool and action scoping, guardrails, containment and rollback, evaluation and regression testing of non-deterministic behaviour, observability, audit traceability and the full agent lifecycle.
• Experience leading a combined AI architecture and engineering organisation in which design authority is separate from build, and governing the move of AI tools and the data fabric from build to operation under a managed service vendor.
• Enterprise AI and ML security experience covering a standards library, AI threat modelling, model and agent risk tiering, adversarial testing and shadow AI discovery. This includes prompt injection, data poisoning, model extraction, excessive agency and AI supply-chain risk, referencing MITRE ATLAS and OWASP LLM Top 10.
• Experience with enterprise data fabric work, from classification architecture to a built and tested pipeline, plus cloud security posture governance across Azure and GCP.
• Track record of leading leaders of teams across several geographies and time zones.
• Executive presence to act as a technical peer to the CISO, CTO and Enterprise AI leadership, and to decline funded initiatives credibly.
• Preferred: senior security certification such as CISSP or CCSP, plus AI security or AI governance credentials.
• Preferred: working knowledge of NIST AI RMF, ISO/IEC 42001, the EU AI Act and the OWASP Top 10 for LLM Applications, applying GxP and data integrity requirements to AI, and knowing where each framework stops being useful.
• Preferred: experience with LangChain or LlamaIndex, vector databases and retrieval tuning, production Python and Kubernetes.
• Preferred: familiarity with MITRE ATLAS and published adversarial ML research.
• Preferred: hands-on experience with LLM orchestration, RAG patterns, vector databases and MLOps tooling.
• Preferred: Certified Kubernetes Security Specialist, or equivalent proven container security depth.
• Preferred: experience in GxP-regulated or other safety-critical environments where an automated action has real-world consequences.
• Preferred: experience in pharmaceutical, life sciences or another highly regulated industry.
• Preferred: experience with enterprise AI programmes and business partnering on use-case intake, feasibility and risk-based prioritisation.
• Preferred: experience building and keeping scarce AI engineering and AI security capability, including growing a technical team.
• Keywords for search: AI for Cybersecurity, Application Security, CI/CD, Cloud Security, Computer Programming, Data Security, Identity Access Management (IAM), Information Systems, Infrastructure Security, Kubernetes, LangChain, LangGraph, Large Language Models (LLMs), LlamaIndex, Machine Learning Operations, MITRE ATT&CK Framework, MLflow, ML Security, Open Web Application Security Project (OWASP), Operating Systems Security, OWASP Top 10, Python, Scripting, Security Engineering, and 3 more.

Conditions

Based in the United Kingdom with a hybrid model: 2 or 3 days a week onsite at the central London headquarters.

Closing date: 6 November 2026.

The company offers an agile working culture for all roles and encourages candidates to ask the hiring team about flexibility.

Equal opportunity: the company does not accept referrals from employment businesses or agencies without prior written authorisation.

Candidates who are US licensed healthcare professionals may have interview-related expenses reported to meet US transparency requirements.