Technical AI Policy Researcher, Frontier Risk - Trust and Safety

The Trust & Safety (T&S) Responsible AI Policy team's mission is to ensure the development of GenAI models and applications are safe, fair and trustworthy. We do this by defining, measuring and mitigating safety and fairness AI model risks through policy frameworks, model risk assessments, and upstream policy solutions.

The T&S Responsible AI Policy team sits within the T&S GenAI and Emerging Products pillar. We work closely with Trust & Safety teams (product policy, product, engineering, data science, operations, red teaming), business and model teams, and cross-functional stakeholders (comms, legal, public policy) across global markets. Success in this team requires strong policy acumen, judgment, creativity, analytical rigour, and the ability to translate Generative AI risk to different stakeholders effectively.

As a Technical AI Policy Researcher on the T&S Responsible AI Policy team, you will champion the responsible development and deployment of our frontier AI models across multiple businesses with a specialty on technical research for frontier risks. You will accelerate technical policy research, incubate new research efforts, and drive end-to-end policy to evaluate workflows for your domain areas.

Responsibilities:
- Design and maintain multimodal GenAI policies across safety-relevant domains, including frontier risk in agentic modalities, loss of control or deceptive misalignment.
- Translate risk and harm models into clear behavioral specifications, evaluation criteria, grading guidance, and system-level safeguards.
- Define practical boundaries between beneficial uses of AI and assistance that could materially enable harm, exploitation, misuse, or unsafe outcomes.
- Build policy artifacts that support model training, evaluation, and deployment. Partner with safety researchers, engineers, product teams, and other stakeholders to operationalize policy into scalable model behavior and measurable safeguards.
- Design end-to-end policy to pre-launch evaluation to post-launch monitoring workflows across safety-relevant domains, including golden set construction, labeling guidance, calibration, adjudication, and eval coverage analysis, to ensure policies can be reliably measured and improved.
- Use red-teaming results, deployment data, model failures, over-refusals, under-refusals, and ambiguous edge cases to improve policy and evaluation quality over time.
- Identify emerging capability areas where frontier AI systems could create new safety, fairness or bias challenges or lower barriers to harm.
- Monitor post-launch model activity to identify gaps in our policy framework to capture unsafe model behaviour.
- Champion research to strengthen the defensibility and operability of policy positions, including working with Outreach and Partnerships to incorporate external expert input into relevant policy positions.
- Combine longer-horizon safety research with hands-on launch and deployment work.
- Contribute to risk reports, policy documentation, launch reviews, and AI governance reviews on the company's approach to building AI responsibly.
- Support regulatory teams as a subject matter expert on AI compliance related initiatives.

Minimum Qualifications:
- 5 years in Trust & Safety, AI Safety Research, AI Ethics, technical AI Governance, or equivalent experience.
- Advanced degree in Computer Science, Human-Computer Interaction, Engineering, Data Science or quantitative Social Sciences
- Direct experience in frontier risk research, AI evaluations, red-teaming, or AI governance work.
- Strong technical understanding of LLM, multimodel, or genmedia model behavior, model failure modes, and safety risks.
- Demonstrated experience working with external experts and stakeholders, including civil society, government, and academia.
- Demonstrated success working in a fast-paced technology company or research organization conducting AI impact, risk assessments or algorithmic audits, and/or data science or product development related experience.
- Ability to advocate for safety amongst a wide variety of business stakeholders including Product Policy, Engineering, Public Policy, Legal, Communications, and Data Science.

Preferred Qualifications:
- Technical knowledge in high efficiency on device architectures, multimodal understanding, V&V of AI systems, or RAG is beneficial but not required.
- Experience working with governments, frontier AI companies, or AI Safety organizations.
- Experience working in non-western cultures, with a particular emphasis on the global south.
- Understanding of Trust & Safety positioning in the entertainment media technology sector, with comfort learning internal tools and product workflows.