AI Safety Scientist (Post-Training & Interpretability)

United States Digital Space LLC; New York, San Francisco (NY, CA); On-site; USD 216,000 - 270,000; Full time

Benefits offered by this job

Health, dental and vision coverage

Equity

Learning and development stipend

Generous PTO

Job summary

Scale Labs in New York is seeking a Research Scientist focused on Safety Post-Training to develop methods and interpretability techniques that make frontier AI systems safer and better understood by researchers and policymakers.

You will design post-training pipelines, develop interpretability-informed evaluations, and collaborate with policymakers, engineers, and researchers to translate findings into tangible safety standards and benchmarks.

Qualifications

• Commitment to safe, secure and trustworthy AI deployments as frontier capabilities advance.
• Experience with post-training and RL techniques such as RLHF, DPO, GRPO, and similar approaches.
• Track record of published research in machine learning, particularly in generative AI.
• 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.

Responsibilities

• Design and run post-training pipelines to study how training choices affect model safety, robustness, and alignment properties.
• Develop interpretability-informed evaluations that reveal how and why models produce unsafe or undesirable behaviors, guiding mitigations.
• Collaborate with policymakers, engineers, and researchers to translate post-training and interpretability findings into actionable safety standards, evaluation benchmarks, and best practices.

Skills

Post-training methods

Interpretability techniques

Policy research

RLHF

DPO

GRPO

Written and verbal communication

Job description

Scale Labs in New York is seeking a Research Scientist focused on Safety Post-Training to develop methods and interpretability techniques that make frontier AI systems safer and better understood by researchers and policymakers.

You will design post-training pipelines, develop interpretability-informed evaluations, and collaborate with policymakers, engineers, and researchers to translate findings into tangible safety standards and benchmarks.