AI Safety Training Researcher — Robustness Focus

OpenAI; United States; On-site; USD 380,000 - 500,000; Full time

Job summary

OpenAI in the United States seeks a researcher to train and evaluate models for U.S. government use, focusing on national security applications. You will advance safety post-training and robustness, helping models follow nuanced policies while preserving usefulness.

Join a collaborative team advancing safety through research, engineering, and policy partnerships to deploy safe, reliable AI systems.

Qualifications

• 4+ years of AI safety research experience (e.g., RLHF, adversarial training, robustness)
• Degree in computer science, machine learning, or a related field
• Strong deep learning research or engineering skills
• Experience deploying AI safety improvements in models or products
• Security clearance or eligibility to obtain TS/SCI where required

Responsibilities

• Research and implement methods for safety training, reinforcement learning, and adversarial robustness.
• Develop evaluations, identify model failure modes, and use findings to improve training.
• Work with research, engineering, security, and policy partners to support safe, reliable deployment

Skills

Reinforcement learning

Adversarial robustness

Safety training methods

Deep learning

Python

PyTorch

Education

Bachelor's degree in CS/ML or related field

Tools

Python

PyTorch

Job description

OpenAI in the United States seeks a researcher to train and evaluate models for U.S. government use, focusing on national security applications. You will advance safety post-training and robustness, helping models follow nuanced policies while preserving usefulness.

Join a collaborative team advancing safety through research, engineering, and policy partnerships to deploy safe, reliable AI systems.