jobs

Research Scientist, Agent Robustness

Scale Labs, Research Scientist — Agent Robustness As the leading data and evaluation partner for frontier AI companies, Scale plays an integral role in understanding the capabilities and safeguarding AI models and systems. Building on this

Company

Scale AI

Location

San Francisco, CA; New York, NY

Status

Open

Posted

2026-03-20T18:48:59+00:00

Scale Labs, Scale AI's policy-research group, seeks a research scientist for agent robustness. Work includes benchmarking agent capabilities against safety and risk factors, building harnesses that pressure agents toward harmful actions, designing exploits and mitigations for failure modes that emerge as agents gain coding, browsing and computer-use abilities, and studying multi-agent risks. Requirements include collaborative technical research, published generative-AI work, three or more years on complex ML problems, and post-training methods such as RLHF, DPO and GRPO. On-site in San Francisco or New York; interviews avoid LeetCode-style questions.