Director, AI Risk Governance & Validation Lead

Responsibilities: 1. Lead the DMO AI Risk Governance & Validation Team • Lead a four-member specialist team covering AI model governance, model validation, AI testing / QA / evaluations, and AI engineering for governance validation. • Set the operating cadence, work allocation, review standards, escalation protocols and evidence-quality expectations for the team. • Build a practical Line 1.5 AI governance capability that supports AI use case intake coordination, preliminary risk triage, evidence completeness checking, control effectiveness challenge and post-deployment governance. • Ensure the team remains a technical assurance and evidence-review function, not an AI delivery owner, model owner or final risk approval owner. • Develop team capabilities across AI governance, model risk, model validation, testing, data governance, GenAI, agentic AI, MLOps and regulatory expectations. 2. Operate Line 1.5 AI Governance Review and Challenge • Review AI use cases from data governance, data risk, model lifecycle and practical control perspectives before escalation to independent risk and compliance stakeholders. • Challenge the completeness and quality of AI review submissions, including business purpose, data sources, model / agent design, preliminary risk classification, validation evidence, control evidence and monitoring plans. • Review whether AI controls are operationally workable, evidenced and sustainable, rather than policy-level only. • Coordinate with Business / Use Case Owners, Data Owners, AI delivery teams, Technology / Security teams, Risk Management Department, Compliance Department, Legal Department, local data protection officers and other relevant control functions. • Provide clear review conclusions, challenge points and remediation recommendations to support independent risk review and governance committee decision-making. 3. Data Governance for AI • Review AI input data sources, data ownership, data classification, personal data / PII treatment, data access controls, cross-border considerations, data quality, lineage, metadata and evidence readiness. • Confirm whether data used for AI use cases is appropriately approved, fit for purpose, traceable and subject to adequate controls. • Challenge whether data access, storage location, output retention, SharePoint / platform access and downstream use are consistent with applicable data governance and privacy requirements. • Work with Data Governance, Data Owners and control functions to establish a unified evidence layer covering data classification, data access, cross-border sharing, data protection, data quality, lineage, metadata and AI data risk assessment. • Support AI-ready data governance standards, evidence templates and review checklists. 4. Model / Agent Validation Challenge • Oversee review and challenge of model / agent validation evidence, including methodology, assumptions, feature logic, data inputs, limitations, performance, stability, robustness, explainability and monitoring design. • Ensure validation evidence is appropriate for use case risk level, intended use and AI lifecycle stage. • Challenge model / agent performance metrics, drift monitoring, bias / fairness assessment, robustness testing, output accuracy testing and human-in-the-loop controls. • Work with model developers, AI engineers, data scientists and independent risk reviewers to ensure validation artefacts are clear, complete and decision-useful. • Ensure validation evidence clearly identifies limitations, residual risks, control gaps and remediation actions where required. 5. AI Testing, QA and Evaluation Oversight • Oversee testing approaches for statistical models, ML models, LLM applications and agentic AI systems. • Ensure coverage of functional testing, regression testing, scenario-based testing, edge cases, adverse / irregular scenarios, bias / fairness evaluation, robustness analysis, drift detection and workflow reliability. • Review end-to-end AI workflows, including data inputs, feature transformations, task completion, tool-use accuracy, API / connector behaviour, multi-step reasoning and output quality. • Promote test logs, evaluation results, benchmarking, traceability and observability evidence to support AI risk review. • Ensure testing findings are documented clearly and translated into remediation actions, risk caveats or acceptance considerations for governance forums. 6. AI Engineering and MLOps Governance • Provide leadership oversight over technical review of AI engineering, agentic workflows, RAG, API integration, tool-use orchestration, Copilot Studio / Power Automate-type workflows, logging and monitoring controls. • Challenge whether AI systems have appropriate MLOps / lifecycle controls, including model registry, version control, deployment controls, monitoring, retirement triggers and change management. • Review whether AI / GenAI solutions disable inappropriate model training or data leakage pathways where required. • Assess whether technical architecture and workflow design support auditability, explainability, resilience and responsible AI expectations. • Partner with Technology, Security and AI delivery teams to embed technical controls early enough in the lifecycle. 7. Governance Framework, Procedures and Evidence Standards • Translate AI governance policy, risk appetite and regulatory expectations into practical review procedures, templates, evidence packs and operating standards. • Maintain AI review checklists and evidence standards covering lifecycle governance, data handling, access control, validation, testing, monitoring and control effectiveness. • Ensure DMO review outputs are structured, audit-ready and reusable for independent risk review and committee escalation. • Develop reporting on review pipeline, common evidence gaps, key control weaknesses, review turnaround, remediation status and recurring AI risk themes. • Support continuous improvement of the AI Risk…