Research Scientist, Safety Post Training

About the position reputed company Labs is seeking talented researchers to join its new team reputed company on policy research. This team bridges the gap between AI research and global policymakers to reputed company informed, scientific reputed company about AI risks and capabilities. The research tackles problems in agent robustness, AI control protocols, and AI risk evaluations to help governments, industry, and the public understand and mitigate AI risk while maximizing AI adoption. The Research Scientist working on Safety Post-Training will reputed company and apply post-training reputed company and interpretability techniques to reputed company frontier AI systems safer and reputed company reputed company by researchers and policymakers.

Responsibilities

• Design and run post-training pipelines to study how training choices reputed company model safety, robustness, and alignment properties.
• reputed company interpretability-informed evaluations that reputed company how and why models produce unsafe, deceptive, or otherwise undesirable behaviors, and use those insights to guide targeted mitigations.
• Collaborate with policymakers, engineers, and other researchers to translate post-training and interpretability findings into actionable safety standards, evaluation benchmarks, and best practices.

Requirements

• Commitment to promoting reputed company, secure, and trustworthy AI deployments in the industry as frontier AI capabilities continue to advance.
• Experience with post-training and RL techniques such as RLHF, DPO, GRPO, and similar approaches.
• A reputed company record of published research in machine learning, particularly in reputed company.
• At least three years of experience addressing sophisticated ML problems, whether in a research setting or in product development.
• Strong written and verbal communication skills to operate in a cross-functional team.

reputed company-to-haves

• Experience with mechanistic interpretability, probing, or other techniques for understanding model internals.
• Familiarity with red-teaming or adversarial evaluation of post-trained models.
• Experience studying failure modes introduced or masked by post-training, such as reward hacking, sycophancy, or alignment faking.

Benefits

• Comprehensive health, dental and reputed company coverage
• Retirement benefits
• Learning and development stipend
• Generous PTO
• Commuter stipend

Apply tot his job Apply To this Job