About this role
Professionals who enjoy fast-paced, collaborative research environments and possess strong judgment regarding the real-world risks of advanced AI will thrive in this position. The Safety Systems team at OpenAI manages the complete lifecycle of safety efforts for frontier models, focusing on system-level safeguards, evaluation, and red-teaming to mitigate misuse and misalignment.
As part of this group, the Model Policy team designs the rules that govern safe model behavior in real-world settings. The incoming Model Policy Manager will shape how OpenAI identifies, evaluates, and addresses risks in multimodal models like GPT-Live and ChatGPT Images. Daily work involves translating threat models into behavioral safety policies, designing evaluation criteria, and analyzing failure patterns across audio, image, video, and omni-modal outputs.
The role exists to ensure that powerful frontier models remain aligned with human values as capabilities expand. Candidates who do well here are pragmatic, evidence-driven, and comfortable turning ambiguous safety questions into clear data-driven policies that can be realistically scaled and supervised.
What you will do
- Design and maintain model policies covering audio, image, video, and omni-modal behavior.
- Translate theories of harm and threat models into behavioral safety policies, evaluation criteria, grading guidance, and safeguards.
- Identify and analyze safety regressions and failure patterns to uncover gaps in existing policies and inform policy iteration.
- Develop policy artifacts that support model training, evaluation, and deployment, including behavior instructions, human-data campaigns, golden sets, and evaluations.
- Partner with AI researchers, domain experts, and product teams to operationalize policy into measurable model behavior.
What you need
- Demonstrate strong judgment concerning the real-world risks of advanced multimodal AI systems.
- Possess experience turning ambiguous safety questions into clear data-driven policies, behavioral boundaries, and measurable evaluation criteria.
- Treat policy as an end-to-end, measurable system by testing whether it produces intended model behavior and diagnosing gaps across policy, data, graders, and safeguards.
- Leverage strong technical judgment to design policies around model behavior that can realistically be trained, measured, and supervised at scale.
- Demonstrate strong technical fluency and use AI tools to accelerate policy development, evaluate model behavior, analyze failure patterns, and turn findings into actionable improvements.
- Work hands-on with model data and evaluation results, including inspecting examples, analyzing failure patterns, assessing data quality, and distinguishing policy failures from grader, model, or system failures.
- Take a pragmatic, evidence-driven approach to reducing risk while preserving beneficial uses of AI.
- Bring hands-on experience driving consensus and action in ambiguous spaces.
Benefits
- Offers Equity
- Relocation assistance to new employees
Skills and tech stack
Questions candidates ask
What is the location and remote work policy for this job?
This role is based in San Francisco, CA. The company uses a hybrid work model of 3 days in the office per week.
What compensation and financial benefits are included?
The salary range is $266K – $335K and the position offers equity. Relocation assistance is also available to new employees.
What will I actually be working on regarding model safety?
The work focuses on the safety of multimodal models like GPT-Live and ChatGPT Images. Responsibilities include designing model policies for audio, image, video, and omni-modal behavior, as well as translating threat models into behavioral safety policies and evaluation criteria.