Research Program Manager, Safety Systems

OpenAI•San Francisco, CA

7d•$239,000 - $328,000•Hybrid

About The Position

OpenAI’s Safety Systems team manages the lifecycle of safety efforts for frontier models, working across research, engineering, product, and policy to understand risks, develop and validate safeguards, and support responsible deployment. Its work includes safeguards research and deployment, safety assessments, launch readiness, and trusted-access programs addressing risks such as biological and cyber misuse and misalignment. Research Program Managers lead programs in AI Safeguards or Frontier Risks Safety. They develop and improve safeguards across models and products, or assess and mitigate risks involving frontier capabilities, launches, and trusted-access programs. They partner with researchers, engineers, and other stakeholders to turn open safety questions into programs of work, shape technical roadmaps, coordinate execution, and use evidence to guide priorities and development and deployment decisions. The role is based in San Francisco, California, with a hybrid work model requiring three days per week in the office. Relocation assistance is offered to new employees.

Requirements

At least 5 years of experience in technical safety work, owning technical programs, or related roles in fast-moving environments.
Experience leading complex research or technical programs involving researchers and engineers, particularly in AI, cybersecurity, biosafety, or another field where failures can have serious consequences.
Understanding of frontier AI safety concepts, including capability evaluations, alignment, control, monitoring, and safeguard robustness, with the ability to engage substantively with technical teams.
Ability to assess what evaluation results do and do not establish, including methodological limitations, uncertainty, and the distinction between model capability and real-world risk.
Ability to turn ambiguous questions into clear programs, make practical tradeoffs, and adjust priorities as evidence changes.
Extensive experience working across research, engineering, and product teams, and coordinating with policy and communications partners.
Sound judgment when findings are concerning or incomplete, including raising unresolved risks constructively and communicating complex technical work clearly.
Motivation to help ensure increasingly capable AI systems remain beneficial and under meaningful human control.

Responsibilities

Partner with research and engineering leads to define safety program goals, milestones, dependencies, and success criteria; prioritize intake and emerging findings, resolve blockers, and balance immediate needs with longer-term research and platform investments.
Translate safety research into safeguards across models and products, track safeguard coverage and effectiveness, identify opportunities for improvement, and drive roadmaps across launches.
Organize capability evaluations, alignment and control assessments, adversarial testing, and safeguard stress-testing, including assessments of autonomous and multi-agent systems throughout model development, internal use, and deployment.
Coordinate safety requirements and execution across frontier model launches and deployments, and work with technical teams to define assessment scope and evidence needed for safety and access decisions.
Triage safeguard gaps, jailbreaks, and overflagging with accountable teams; establish owners and escalation paths; track remediation and regression testing; and turn recurring issues into research and engineering improvements.
Support safety cases connecting claims to evaluations, safeguards, assumptions, and remaining uncertainty; identify evidence gaps and unresolved risks; and communicate methods, results, limitations, and follow-up actions through assessment reports, system cards, and decision-maker updates.
Coordinate shared tooling and feedback workflows, privacy and security requirements, safeguard user experience, and quality assurance to help teams apply safeguards consistently across products and launches.