Research Scientist (AI)

Salary: $216,000–$270,000 USD per year, with potential equity and benefits.

Company

Scale AI develops data, evaluation, and full-stack technologies that support reliable AI systems and help organizations build, deploy, and oversee AI applications.

What you will do

• Design and run post-training pipelines to study how training choices affect model safety, robustness, and alignment.
• Develop interpretability-informed evaluations for unsafe, deceptive, and otherwise undesirable model behaviors.
• Use evaluation and interpretability insights to guide targeted safety mitigations.
• Collaborate with policymakers, engineers, and researchers to translate findings into safety standards, benchmarks, and best practices.

Requirements

• Experience with post-training and reinforcement learning methods such as RLHF, DPO, GRPO, or similar approaches.
• Published research in machine learning, particularly generative AI.
• At least three years of experience solving sophisticated machine learning problems in research or product development.
• Strong written and verbal communication skills for cross-functional collaboration.
• Commitment to safe, secure, and trustworthy AI deployment.

Nice to have

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

Culture & Benefits

• Work alongside industry, public-sector, and academic collaborators, with opportunities to publish research.
• Comprehensive health, dental, and vision coverage.
• Retirement benefits, learning and development stipend, and generous paid time off.
• Potential commuter stipend and equity compensation for eligible roles.
• Inclusive equal-opportunity workplace with reasonable accommodation support.

Hiring process

• Research interviews assess practical machine learning prototyping and debugging skills, research concepts, and organizational alignment.
• No LeetCode-style questions are used.
• Applicants are reconsidered for the same role only after a 90-day waiting period.