AI Technical Risk Assurance Manager - Responsible AI & Resilience

Job Titile: AI Technical Risk Assurance Manager

Location: Bangalore

ABOUT UNILEVER
With 3.4 billion people in over 190 countries using our products every day, Unilever is a business that makes a real impact on the world. Work on brands that are loved and improve the lives of our consumers and the communities around us. We are driven by our purpose: to make sustainable living commonplace, and it is our belief that doing business the right way drives superior performance. At the heart of what we do is our people – we believe that when our people work with purpose, we will create a better business and a better world.
At Unilever, your career will be a unique journey, grounded in our inclusive, collaborative, and flexible working environment. We don’t believe in the ‘one size fits all’ approach and instead we will equip you with the tools you need to shape your own future.

ABOUT GDT:

GDT is the global digital and technology engine of Unilever offering business services, technology, and enterprise solutions. GDT serves over 190 locations and through a network of specialized service lines and partners delivers insights and innovations, user experiences and end-to-end seamless delivery making Unilever Purpose Led and Future Fit. 

About the Role

Unilever is scaling AI at extraordinary speed — 191 use cases a year, ~3 a day, spanning 190+ markets, with agentic systems now making autonomous decisions at enterprise scale. With 1 in 4 use cases inherently high-risk, the question is no longer whether AI is being deployed — it is whether it is behaving the way it should.

This role owns that question. As AI Technical Risk Assurance Manager, you are the function responsible for ensuring Unilever's AI systems — across ML, GenAI, and Agentic — are safe, fair, robust, and compliant before they go live and throughout their operational lifecycle. You will be the primary assurance control across the full AI estate: running adversarial testing and red teaming, conducting quantitative and qualitative risk assessments, assuring agentic systems where failure consequences are amplified, and maintaining the evidence stack that regulators, auditors, and senior leadership rely on.

With the EU AI Act now enforceable and agentic AI volumes scaling 10× in 2026, assurance is no longer a checkpoint — it is the control that makes innovation defensible. This role sits at the heart of Unilever's responsible AI ambition: enabling the business to move fast, without breaking trust.

Key Responsibilities

• Lead AI Red Teaming. Design and execute adversarial testing against AI models and agentic systems — simulating real-world misuse scenarios across attack vectors such as jailbreaks, prompt injection, brand impersonation, data leakage, and transactional fraud. Convert abstract risk into concrete, evidence-backed findings.

• Conduct Risk Assessments. Run quantitative and qualitative assessments to validate AI systems against Unilever's Responsible AI standards, the EU AI Act, GDPR, NIST AI RMF, and ISO 42001 — producing defensible evidence for regulators, internal audit, and the risk board.

• Deliver Continuous Re-Assurance. Move beyond point-in-time reviews. Monitor in-production AI systems for model drift, performance decay, and emergent risk — ensuring assurance state reflects reality, not last self-report.

• Assure Agentic AI. Design and execute assurance for autonomous, tool-using, and multi-agent systems. Trace agent actions, detect anomalies, probe for tool misuse and injection — where the blast radius of failure is enterprise-wide.

• Own the Evidence Stack. Maintain mitigation tracking, assurance records, and reporting as the single source of truth for executives, regulators, and audit — from inherent risk at intake to residual risk post-mitigation.

• Drive Left-Shift Assurance. Embed assurance earlier in the AI delivery lifecycle — reducing downstream risk, improving speed to market, and moving Unilever toward a responsible-by-design model.

• Orchestrate Across Functions. Partner with Privacy, Cyber, Legal, Enterprise Architecture, Procurement, and Product teams to embed assurance controls into AI delivery — ensuring governance is an accelerator, not a gate.

Candidate Profile

• 10 years of experience in AI/ML risk, model governance, technical audit, information security, or a closely adjacent field — with a track record of operating in complex, large-scale environments.

• Strong grounding in AI and ML. Solid conceptual and applied understanding of how large language models, generative AI, and agentic systems work — including their architectural failure modes, bias vectors, and emergent behaviours. You know what can go wrong, and why.

• Familiarity with AI red teaming — the methodology, the attack surface, and how it differs from traditional cyber red teaming. Hands-on experience is a strong plus; what matters most is that you understand the space deeply and can learn, adapt, and lead in it.

• Regulatory and framework literacy. Working knowledge of the EU AI Act, GDPR, NIST AI RMF, and ISO 42001 — and the ability to translate regulatory intent into operational assurance controls.

• Evidence and risk mindset. You think in outcomes, not activities — and you know how to build an evidence stack that holds up to scrutiny from regulators, auditors, and boards.

• Sharp communicator. Able to translate complex technical findings into executive-level insight — whether in a risk memo, a board briefing, or a cross-functional conversation with Privacy, Cyber, or Legal.

• Fast learner and go-getter. The AI risk landscape evolves rapidly — new architectures, new attack surfaces, new regulations. You are intellectually curious, self-directed, and energised by ambiguity. You don't wait for a playbook; you help write it.

• Automation-first mindset. You design for scale — not one-off reviews. You look for ways to make assurance faster, more consistent, and embedded by default.

• Collaborative by nature. You work fluidly across functions, build trust with technical and non-technical stakeholders alike, and can influence without authority.

Why This Role Matters

One in four AI use cases at Unilever is inherently high-risk. A biased model, a hallucinating agent, or an undetected drift event can trigger regulatory fines, market withdrawals, litigation, and lasting brand damage — across 190+ markets simultaneously.

This role is the control that makes Unilever's AI-first ambition defensible. It is not a compliance checkbox — it is the function that ensures AI performs the way it should, every time, at scale. You will work on real systems, with real consequences, at a company scaling AI faster than most in the world. For the right person, this is not just a role — it is a defining career opportunity at the frontier of responsible AI.

LEADERSHIP SKILLS

• CARE DEEPLY: We care about how consumers experience our brands, the growth and development of our people, and their impact on the planet. We emphasize the importance of performance and care, moving from ambiguity about success to fairness and transparency. 

• FOCUS ON WHAT COUNTS: We prioritize what truly matters, setting clear and stretching goals. We aim to shift from having everything as a priority to focusing on fewer, bigger things that are delivered to conclusion and are being rewarded. 

• STAY THREE STEPS AHEAD: We encourage bold and creative thinking to make breakthroughs in performance. We focus on anticipating and staying ahead of consumer needs and external trends, shifting from reacting to leading, shaping, and disrupting the market. 

• DELIVER WITH EXCELLENCE: The emphasis is on delivering everything with excellence and pace, taking personal ownership, and holding each other accountable. We aim to shift from pride in thinking to pride in execution, developing breakthrough solutions and ensuring the best outcomes. 

Our commitment to Equality, Diversity & Inclusion 

• Unilever embraces diversity and encourages applicants from all walks of life! This means giving full and fair consideration to all applicants and continuing development of all employees regardless of age, disability, gender reassignment, race, religion or belief, sex, sexual orientation, marriage and civil partnership, and pregnancy and maternity. 

#LI-Onsite