AI Safety Training Researcher – National Security (Remote)

OpenAI; San Francisco (CA); On-site; USD 380,000 - 500,000; Full time

Benefits offered by this job

Equity

Remote work option

Job summary

OpenAI is seeking a Researcher in Safety Training for National Security to advance post-training safety, harness robust policies, and ensure useful behavior in high-stakes government contexts.

You will train and evaluate models for U.S. government use, collaborating with research, engineering, security, and policy partners to deploy safe, reliable systems while advancing safety post-training and robustness.

Qualifications

• 4+ years of AI safety research experience, including RLHF or robustness work.
• Strong deep learning research or engineering skills.
• Experience improving model safety for deployment.
• Aligned with OpenAI’s mission and responsible AI use.

Responsibilities

• Research and implement methods for safety training and robustness.
• Develop evaluations and identify model failure modes to improve training.
• Collaborate with research, engineering, security, and policy partners to support safe deployment.

Skills

AI safety research

RLHF

Adversarial training

Robustness

Education

Degree in computer science, machine learning, or related field

Job description

OpenAI is seeking a Researcher in Safety Training for National Security to advance post-training safety, harness robust policies, and ensure useful behavior in high-stakes government contexts.

You will train and evaluate models for U.S. government use, collaborating with research, engineering, security, and policy partners to deploy safe, reliable systems while advancing safety post-training and robustness.