National Security Safety Researcher

Job Summary

Seeking a full-time National Security Safety Researcher, this position focuses on training and evaluating AI models for U.S. government applications, advancing safety post-training and robustness while collaborating with cross-functional teams in a remote environment.

Key responsibilities

• Research and implement methods for safety training, reinforcement learning, and adversarial robustness
• Develop evaluations, identify model failure modes, and utilize findings to enhance training processes
• Collaborate with research, engineering, security, and policy partners to ensure safe and reliable deployment of models

Required qualifications

• 4+ years of relevant AI safety research experience, including reinforcement learning from human feedback (RLHF) and adversarial training
• Degree in computer science, machine learning, or a related field with strong deep learning research or engineering skills
• Experience in improving model safety for deployment and a passion for collaborative research
• Active TS/SCI clearance or equivalent
• Motivation aligned with the responsible use of AI in safety-critical settings