Role purpose
Infinitas Learning is building an AI Centre of Excellence to scale AI safely and responsibly across our portfolio of digital learning products and internal initiatives. The Portfolio Manager — AI Governance & Live Oversight is the single point of accountability for governance across this portfolio.
In the early phases, you will control distribution of AI work by operating the gating process (ideation experimentation pilot scale), ensuring every initiative meets COE standards before progressing. As the program matures, your remit will expand to live governance of production AI systems across Operating Companies (OpCos) and Group functions — with a focus on ethics, model risk, hallucination mitigation, incident management, and ongoing compliance.
This role combines governance, risk, and program management with a practical understanding of how modern AI systems (LLMs, RAG, agentic patterns) behave in production.
Key responsibilities
-
Gatekeeper for portfolio distribution (early phase)
-
Operate and refine the gating framework (ideation / experimentation / pilot / scale), including submission templates, minimum evidentiary requirements, and approval workflows.
-
Define clear approval criteria for each gate (data readiness, pedagogy/education impact, security/privacy, vendor suitability, model risk level, autonomy level for agentic proposals).
-
Chair the Gate Review panel for incoming proposals; coordinate technical reviewers (Platform/MLOps), Legal, Security, Product, and OpCo representatives.
-
Maintain a portfolio backlog and prioritisation rules to control the flow of initiatives into limited COE/platform capacity, avoid duplication, and enforce reuse of existing capabilities.
2. Model & agentic risk taxonomy, red-team & sandbox coordination
-
Maintain and operationalise the model risk taxonomy and autonomy levels for agentic AI, mapping these to review depth, approvals, and red-team obligations.
-
Coordinate use of the COE sandbox / red‑team process for high‑risk or agentic pilots; ensure behavioural tests and emergent‑behaviour mitigations are passed before production use.
-
Verify that required controls (HITL checkpoints, kill switches, audit trails, rate limits, escalation paths) are in place before authorising production.
3. Live governance & ethical oversight (scale phase)
-
Own live governance for production AI systems: maintain current risk assessments and ensure ongoing monitoring for drift, bias, safety, and hallucination.
-
Design and embed hallucination detection & mitigation frameworks (test suites, monitoring signals, escalation paths) and ensure remediation SLAs are defined and met.
-
Establish and maintain incident response & escalation playbooks for model failures, hallucinations, privacy incidents, or agentic misbehaviour; coordinate cross‑functional post‑mortems and drive lessons learned into standards and controls.
4. Governance tooling, controls & automation
-
Champion and operationalise a governance tooling stack (model registry, approvals/workflow automation, lineage, prompt/version control, monitoring dashboards) that supports both gate reviews and live governance.
-
Work with AI Engineering, Data, and Platform teams to automate compliance checks and accelerate first‑pass reviews without compromising standards.
5. Standards, policies & vendor governance
-
Maintain and update core governance artefacts: AI Governance Framework, Agentic AI Policy, production readiness checklist, and vendor evaluation templates.
-
Review vendor risk and contractual clauses (data processing, IP, explainability, audit, termination) in coordination with Procurement and Legal.
-
Ensure OpCos and Group functions are aware of, and working to, consistent standards for internal builds and vendor solutions.
6. Community coordination & enablement
-
Act as the governance focal point for the AI Community of Practice:
-
train OpCo AI Specialists and Group teams on gate requirements,
-
run onboarding for new initiatives,
-
publish regular governance guidance, FAQs, and playbook updates.
-
Run regular governance reviews with OpCo AI Specialists and AI initiative owners to confirm that live initiatives conform to enterprise standards and to harvest reusable mitigations and patterns.
7. Measurement & reporting
-
Maintain a governance dashboard, including (for example):
-
% projects passing first‑pass review,
-
time in gate,
-
incidents by severity,
-
MTTR for hallucination / safety incidents,
-
% production systems under live governance,
-
% OpCo compliance to core standards.
-
Report portfolio compliance, incidents, and overall risk posture to the Director of AI / Head of AI COE, the AI Steering Group, and Transformation leadership.
Key relationships
-
Executive: Director of AI / Head of AI Centre of Excellence (AI COE), Transformation Sponsor, AI Steering Group.
-
Functional partners: Legal, Security, Procurement, HR, Finance, ILPT, OpCo Leadership and AI Specialists.
-
External: red‑team / security vendors, governance tooling vendors, advisory partners.
-
Day‑to‑day collaboration: AI Engineering Lead, Enablement Lead, Data Governance Lead, Data Analytics Lead, Legal, Security, Procurement, OpCo AI Specialists, Product Managers, and Transformation/TMO.
In future, this role may lead or be supported by a Governance Coordinator / Analyst, depending on scale.
Essential
-
6+ years experience in governance, risk, compliance, security, or AI policy roles, with direct experience in ML/AI governance or model risk management.
-
Practical knowledge of LLMs, RAG, agentic AI concepts and LLMOps/ML Ops practices (production readiness, monitoring, model versioning, rollback strategies).
-
Experience designing and running gating/approval processes, red‑team tests, and sandbox validation for high‑risk or regulated technology initiatives (ideally AI).
-
Strong stakeholder management in matrixed, multi‑country organisations; proven ability to influence Product, Legal, Security, and OpCo leaders.
-
Solid understanding of data protection (e.g. GDPR), privacy by design, accessibility considerations, and procurement risk controls.
-
Excellent program/portfolio management skills and hands‑on experience with governance tooling (model registries, workflow automation, monitoring / risk dashboards).
Desirable
-
Prior experience in education, publishing, or edtech (advantageous).
-
Familiarity with works council consultation and multi‑jurisdictional labour considerations.
-
Certifications in AI governance, security, or risk management (e.g. model risk, information security, data protection).
Personal attributes
-
Meticulous & process‑oriented – strong discipline for checklists, documentation, and evidentiary requirements.
-
Pragmatic – balances risk mitigation with enabling innovation and speed; knows when to say “not yet” and when to enable controlled experimentation.
-
Decisive and assertive – confident chairing gate reviews, challenging assumptions, and making clear approvals, rejections, or conditional approvals.
-
Collaborative – builds trust across OpCos and Group functions; able to hold high standards without creating unnecessary friction.
-
Curious and forward‑looking – keeps up to date with AI safety, hallucination detection, evaluation, and governance tooling, and translates emerging practice into Infinitas’ frameworks.
