AI Governance

ABOUT:

Makes AI systems trustworthy and compliant — governance frameworks, model risk assessment, and regulatory implementation that keep delivery out of legal and reputational trouble.

KEY RESPONSIBILITIES

• Design and implement AI governance frameworks and policy for client engagements
• Conduct model risk assessments and responsible-AI reviews before production go-live
• Own AI regulatory implementation — mapping engagement requirements to relevant regulation (EU AI Act, sector rules)
• Design and run AI red-teaming exercises for high-risk systems
• Own AI testing & evaluation standards from a governance/ethics lens (bias, fairness, explainability, AI ethics)
• Advise practice leads and Architects on governance risk during solution design, not just at the end
• Facilitate cross-functional governance reviews across legal, technical, and business stakeholders
• Be the voice that slows down a launch when risk warrants it, under commercial pressure
• Train delivery teams on responsible-AI practices

REQUIREMENTS & SKILLS

• 6–12+ yrs in AI governance, model risk, compliance, or responsible-AI roles
• Working knowledge of AI regulation (EU AI Act, NIST AI RMF, sector-specific frameworks) and how to operationalize it
• Technical enough to critically assess a model card or evaluation report, not just check a box
• Familiarity with bias/fairness testing methodologies and explainability tooling
• Familiar with platform-native AI safety/governance tooling — Microsoft Azure AI Content Safety, AWS Bedrock Guardrails, and Google Vertex AI safety filters; open-source evaluation tooling (Giskard, DeepEval) a good-to-have
• Experience designing red-team protocols for AI systems
• Strong written policy/framework authorship — produces documents legal and technical teams both trust
• Comfortable being the voice that slows down a launch when risk warrants it, under commercial pressure
• Facilitates cross-functional governance reviews across legal, technical, and business stakeholders
• Diplomatic but firm — holds the line on risk without becoming an obstructionist