Skip to content

Researcher, Frontier Robustness

OpenAISan FranciscoRemote

Estimated salary, not published by the employer: Est. $102k–153k/yr

Needs a free Superjobs account. You'll finish on OpenAI's Ashby page.

Skill score

Summary

This researcher designs experiments, threat models, and adversarial tests to find and fix failure modes in sandboxes, research infrastructure, and model safeguards for frontier AI systems. It suits someone with hands-on LLM research experience and security judgment who can work through open-ended problems with Security and infrastructure partners.

Estimated salary
Est. $102k–153k/yr
Estimated for a mid-level role, ~3 yrs experience (typical for the level), United States. Not published by the employer.
Work mode and location
Remote, San Francisco
Employment type
Full-time

Job description

Overview

The Frontier Robustness team advances the science of securing frontier AI systems. We develop new methods to keep increasingly capable models contained, observable, and under human control, focusing on the sandboxes, shared research infrastructure, and protective systems used during model development and evaluation.

We use our most capable models to discover and combine weaknesses across the security stack, studying how they might escape containment, communicate through unauthorized channels, or evade monitoring. We turn these findings into new defenses and test whether protections hold as models, environments, and attack strategies change.

We are looking for a researcher to develop and test new methods for securing frontier AI systems. You will investigate open-ended questions about model behavior, containment, and safeguards—designing experiments, building evaluations, and developing defenses where established approaches may be insufficient. You will work closely with the Security organization to define the problems and translate research findings into practical protections.

You will take research from an initial question and threat model through experiment design, reproducible findings, and tested defenses. You will write code, build tools and prototypes, and work closely with researchers and infrastructure owners to turn new ideas into safeguards that work in practice. This requires sound security judgment, comfort with uncertainty, and the ability to devise solutions to problems without established answers. This role is based in San Francisco.

Roles and responsibilities

  • Design and run experiments using frontier models and hands-on security techniques to discover failure modes in sandboxes, research infrastructure, and the services models can reach. Reproduce vulnerabilities and assess their impact.
  • Develop threat models and novel adversarial tests for attacks that cross multiple security boundaries, including sandbox escapes, unauthorized communication between agents, and access to sensitive data or systems through shared services.
  • Invent, prototype, and evaluate defenses that strengthen isolation and limit the impact of a compromised boundary. Study agent permissions and tool configurations to identify excessive access and build layered protections with infrastructure owners.
  • Develop and red-team model-based safeguards and monitors of model reasoning and actions. Investigate evasion, test new monitoring techniques, and measure when protections detect or prevent unsafe behavior.
  • Develop automated adversarial tests and repeatable evaluations to measure progress as model capabilities, configurations, and attack strategies change. Assess whether defenses generalize to new attacks, verify fixes, and identify regressions.
  • Partner across Research, Security, and infrastructure teams, including Agent Security, to prioritize risks and turn findings into deployed improvements while accounting for performance and researcher workflows.

Skills required

Top skills

  • Python programming
  • threat modeling
  • experiment design
  • reproducible evaluations

All skills (9)

  • large language models
  • operating systems
  • containers and virtualization
  • distributed infrastructure
  • AI agents
  • sandboxing
  • vulnerability research
  • adversarial testing
  • monitoring and detection

Qualifications

  • Have hands-on experience conducting research with large language models, including designing and running experiments and evaluating model behavior.
  • Can turn an open-ended question into a clear hypothesis, build reproducible experiments at scale, and use the results to develop or improve a method.
  • Can design rigorous evaluations, investigate unexpected results, and test whether findings hold across different models, environments, and attack strategies.
  • Write and debug code independently, build research tools and prototypes, and carry ideas through experimentation.
  • Bring practical security experience relevant to this work, such as sandboxing, vulnerability research, adversarial testing, or detection and monitoring.
  • Work effectively with Security and infrastructure partners to develop threat models, interpret findings, and turn promising research into deployed safeguards.

Nice to have

  • Experience with operating systems, containers or virtualization, distributed infrastructure, or AI agents would also be valuable.

Company info

OpenAI is an AI research and deployment company that builds AI systems and deploys them to the world through its products. Its customers and businesses use these AI products across their operations, with support from its sales, solutions, and support teams.

More jobs at OpenAI