About the role
We're looking for a Forward Deployed Risk Expert to embed directly within our AI delivery squads and make sure risk, governance and assurance are built into how we ship AI — not bolted on at the end. You'll be the connective tissue between business, technology, product, risk, compliance, security and independent testing, turning ambiguous governance expectations into practical controls, testable requirements and decision-ready evidence.
This is a delivery-embedded role, not a back-office review role. You'll sit with the squad from Sprint 0 through production, challenge testing evidence, prepare governance submissions, and move decisions forward across teams that don't report to you.
What you'll do
- Act as the embedded AI risk partner for high-risk AI/GenAI use cases across the full lifecycle; intake, design, build, test, deploy and monitor.
- Translate AI risk, regulatory and policy expectations into delivery artefacts: risk registers, control/guardrail designs, test strategies, acceptance criteria and evidence packs.
- Design and challenge risk-led testing so evidence is traceable, auditable and decision ready for governance teams
- Review testing evidence (golden datasets, adversarial/red-team, LLM-as-judge, manual review, regression) and flag gaps in coverage, traceability or quality.
- Own assurance process integrity while delivery, business, security and risk owners retain their accountabilities.
- Prepare testing summary for risk assessment, operational guardrails and release-readiness recommendations, including conditional-approval positions.
- Coordinate governance submissions and decision packs for senior stakeholders and governance forums.
- Define post-release monitoring, observability, alerting thresholds and incident-response expectations before deployment.
- Build reusable templates, playbooks and standards to scale AI governance beyond one-off reviews.
What you'll deliver
Risk classifications · risk registers & control mappings · guardrail designs · risk-led test methodologies · scenario/test-case catalogues · evidence workbooks · residual risk & operational guardrail papers · governance submission packs · sign-off trackers · post-release monitoring plans.
What you'll bring
Required
- 10+ years across AI governance, model/technology/data risk, compliance, assurance or AI/data delivery.
- Experience delivering AI, ML, GenAI or data platforms in a regulated environment (financial services/insurance preferred).
- Track record creating or reviewing risk documentation, test evidence, control assessments or audit-ready artefacts.
- Practical understanding of the AI lifecycle: intake, design, testing, validation, deployment and monitoring.
- Proven ability to influence cross-functional teams without formal authority.
- Strong communication with both technical and non-technical stakeholders.
Preferred
- Hands-on exposure to GenAI/LLM applications, RAG evaluation, red teaming or human-in-the-loop review.
- Familiarity with MLOps, CI/CD, Databricks/MLflow or cloud AI platforms.
- Knowledge of MAS / APAC regulatory expectations for AI, data protection, technology and conduct risk.
- Familiarity with NIST AI RMF, ISO AI standards, model risk management or responsible AI frameworks.
How we'll know you're succeeding
- Risks are surfaced early and linked to controls, tests and evidence.
- Delivery teams know what evidence is needed before governance review.
- Residual risks are explicitly documented, owned, accepted, remediated or escalated.
- Governance submissions are decision-ready — not reliant on verbal alignment.
- Governance becomes embedded in delivery rather than a late-stage bottleneck.
Key stakeholders
Business & product owners · data scientists & ML/software engineers · solution architects · QA/UAT & independent testing · technology risk, security & architecture · legal, compliance & conduct · AI governance forums · senior leadership.
