AI Safety & Agent Robustness Scientist

AISafety; New York, Northern (NY, KY); Hybrid; USD 216,000 - 270,000; Full time

Benefits offered by this job

Health insurance

Dental coverage

Vision coverage

Generous PTO

Job summary

Scale Labs is seeking a Research Scientist focused on agent robustness to advance safe and aligned AI systems. You will study agent capabilities, develop evaluation harnesses, and prototype mitigations for complex failure modes across interacting AI agents.

This full-time role offers opportunities to publish, collaborate across industry and academia, and contribute to safety protocols as frontier AI capabilities evolve.

Qualifications

• 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

• Research the science of AI agent capabilities with a focus on safety, risk factors, and benchmarking methodologies.
• Design and build harnesses to test AI agents’ tendency to take harmful actions when pressured or manipulated by the environment.
• Design and build exploits and mitigations for new failure modes as AI agents gain capabilities like coding, web browsing, and computer use.
• Characterize and design mitigations for potential failure modes or broader risks in systems with multiple interacting AI agents.

Skills

RL techniques

Agent scaffolding

Evaluation harnesses

Cross-functional collaboration

Publication experience

Tools

SWE-bench

WebArena

OSWorld

Inspect

Job description

Scale Labs is seeking a Research Scientist focused on agent robustness to advance safe and aligned AI systems. You will study agent capabilities, develop evaluation harnesses, and prototype mitigations for complex failure modes across interacting AI agents.

This full-time role offers opportunities to publish, collaborate across industry and academia, and contribute to safety protocols as frontier AI capabilities evolve.