Frontier AI Safety Researcher | Mitigations & Alignment

OpenAI; San Francisco (CA); On-site; USD 295,000 - 445,000; Full time

Job summary

OpenAI seeks exceptional researchers to push the frontier of safety mitigations for frontier models. You will help derisk models by developing novel mitigations, applying methods from interpretability, control, and alignment to ensure deployed models remain safe.

You will collaborate with experts across misalignment, cybersecurity, and biology to build a scalable safety stack and define how a safe AI system should perform in the future at OpenAI.

Qualifications

• 2+ years of experience in AI safety or related field.
• PhD or equivalent required for the role.
• Experience with large-scale AI systems and safety practices is preferred.

Responsibilities

• Identify emerging AI safety risks and develop mitigation strategies.
• Build and refine evaluations to measure risk and safety performance.
• Set research directions to improve safety, alignment, and robustness of AI systems.
• Contribute to best practices guidelines for AI safety across OpenAI and the industry.
• Design and evaluate red-teaming pipelines to test end-to-end safety.

Skills

Python

RLHF

Interpretability

Safety

Control

Robustness

Education

PhD in Computer Science / ML or related field

Job description

OpenAI seeks exceptional researchers to push the frontier of safety mitigations for frontier models. You will help derisk models by developing novel mitigations, applying methods from interpretability, control, and alignment to ensure deployed models remain safe.

You will collaborate with experts across misalignment, cybersecurity, and biology to build a scalable safety stack and define how a safe AI system should perform in the future at OpenAI.