Researcher, Frontier Robustness

OpenAI•San Francisco, CA
•$295,000 - $500,000•Onsite

About The Position

The Frontier Robustness team advances research into securing frontier AI systems by developing methods to keep capable models contained, observable, and under human control. The team examines sandboxes, shared research infrastructure, and protective systems used during model development and evaluation. It uses frontier models to investigate weaknesses in security systems, including potential escapes from containment, unauthorized communication, and evasion of monitoring, and uses the findings to develop and test defenses. The Researcher will develop and test methods for securing frontier AI systems, investigating open-ended questions about model behavior, containment, and safeguards. The role involves designing experiments, building evaluations, and developing defenses, including where established approaches may be insufficient. The researcher will work with the Security organization and infrastructure owners to define problems and translate research findings into practical protections. The role takes research from an initial question and threat model through experiment design, reproducible findings, and tested defenses. It includes writing code, building tools and prototypes, and collaborating with researchers and infrastructure owners to implement safeguards. The position is based in San Francisco.

Requirements

  • Hands-on experience conducting research with large language models, including designing and running experiments and evaluating model behavior.
  • Ability to turn open-ended questions into clear hypotheses, build reproducible experiments at scale, and use results to develop or improve methods.
  • Ability to design rigorous evaluations, investigate unexpected results, and test whether findings hold across different models, environments, and attack strategies.
  • Ability to write and debug code independently, build research tools and prototypes, and carry ideas through experimentation.
  • Practical security experience relevant to the work, such as sandboxing, vulnerability research, adversarial testing, or detection and monitoring.
  • Ability to work effectively with Security and infrastructure partners to develop threat models, interpret findings, and turn research into deployed safeguards.

Nice To Haves

  • Experience with operating systems, containers or virtualization, distributed infrastructure, or AI agents.

Responsibilities

  • Design and run experiments using frontier models and hands-on security techniques to discover failure modes in sandboxes, research infrastructure, and services accessible to models; reproduce vulnerabilities and assess their impact.
  • Develop threat models and 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 monitoring techniques, and measure whether 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.
  • Turn research questions into hypotheses, design reproducible experiments, investigate results, and develop or improve methods.
  • Write and debug code, build research tools and prototypes, and carry ideas through experimentation.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service