Research Program Manager, Safety Measurement

Description

The safety systems group runs the full lifecycle of safety work on the company's frontier models, covering safeguards, model training, evaluation and red-teaming, with the aim of limiting misuse and keeping safety standards high.

The goal is to learn continuously from real deployments and keep powerful tools aligned with human values and safety needs.

This role is based in San Francisco, CA, and relocation assistance is available.

You will work with researchers, data scientists, policy owners and product teams who build on or rely on the measurement systems, helping decide which problems matter most and shaping the data, dashboards, tools and workflows they use.

The role blends measurement expertise, product judgment and cross-functional delivery: you will work with evaluation results and data pipelines, weigh prioritization trade-offs, and take work from problem definition through rollout and iteration.

Responsibilities

• Set strategic priorities and product roadmaps to improve safety measurement and evaluation methods, and extend the measurement platform to new use cases, products and cross-functional needs.
• Work with research, data science, engineering, policy and other stakeholders to define a view of safety outcomes and prevalence across products, and turn it into concrete programs of work.
• Lead the integration of current AI safety research into the company's measurement and evaluation platforms, handling technical trade-offs, dependencies and blockers with researchers, data scientists and engineers.
• Write and keep refining clear, actionable success criteria for safety measurement and evals.
• Report quantitative safety progress to senior leadership.

Requirements

• You have 6+ years in research program management, technical program management, product management or a related role, with experience in AI safety, trust and safety, integrity or a related area.
• You have a solid background in data science, statistics, economics or a related field, and hands-on work with measurement or evaluations.
• You like working where data science meets AI safety research, turning changing measurement needs into clear requirements, priorities and roadmaps.
• You can run complex cross-functional programs on your own, from problem definition to rollout, creating alignment and unblocking decisions in unclear situations.
• You ask sharp questions, test assumptions, and judge technical trade-offs and the strengths and limits of metrics and evaluations well.
• You build strong relationships across research, data science, engineering, policy and product, and explain complex findings clearly to technical teams and leadership.
• You are curious about AI safety and its effects on society, pick up new domains quickly, and adjust as research and priorities change.

Conditions

The role is based in San Francisco, CA.

Relocation assistance is available.

The company is an equal opportunity employer and does not discriminate on the basis of race, religion, color, national origin, sex, sexual orientation, age, veteran status, disability, genetic information or other legally protected characteristics.

Background checks follow applicable law, and candidates with arrest or conviction records are considered in line with San Francisco, Los Angeles County and California fair chance rules.

Applicants who believe a posting is non-compliant can report it through the company's form, and accommodation requests are handled on request.