Responsibilities

As a Novel Abuse Red Teamer, you’ll advance our pre- and post-deployment testing efforts at OpenAI, ensuring our models and products remain resilient against evolving, real-world adversarial threats.
We are seeking a self-starter to design, execute, and refine innovative adversarial testing and simulation protocols that leverage state-of-the-art tools and a hands-on red team approach. In this role, you will simulate threat actor methodologies and conduct rigorous application testing to uncover novel abuse vectors.
We value professionals with excellent communication skills, a commitment to continuous learning, and a passion for securing AI technologies through innovative testing methodologies.
This role is based in San Francisco, CA. We use a hybrid work model of 3 days in the office per week and offer relocation assistance to new employees.
Design and execute high-fidelity, realistic attack simulations that emulate sophisticated threat actor methodologies, uncovering hidden exploitation vectors through advanced red teaming and application testing techniques.
Analyze testing outcomes and generate actionable insights to refine our detection mechanisms, ultimately bolstering our overall security posture.
Serve as a crucial technical bridge between our security, product, and policy teams, ensuring that insights from adversarial testing inform broader defense strategies and remediation efforts.
Develop and integrate automation tools to streamline testing workflows and enable rapid iteration in response to emerging threats.
Collaborate closely with internal stakeholders to align testing strategies with evolving product needs and security objectives.
You might thrive in this role if you have / are:
Experience in application security, penetration testing, or red team operations, with a proven track record of executing adversarial testing and simulation protocols.
A strong technical background in web application security, coupled with hands-on expertise performing security testing of web applications, AI systems, and highly unique attack surfaces.

Novel Abuse Red Teamer at OpenAI · LumiJobs

Novel Abuse Red Teamer

Responsibilities

Qualification

Required

Benefits

Company

OpenAI

More roles at OpenAI