Research Scientist, Open Benchmarks Red Team

Snorkel AI•San Francisco, CA
•$198,000 - $340,000

About The Position

Snorkel AI is hiring researchers to support its Open Benchmarks Red Team program. This hands-on research role combines AI evaluation, adversarial testing, practical benchmark repairs, and collaboration with benchmark authors. The role focuses on identifying weaknesses in frontier AI benchmarks, developing and validating fixes, and creating reusable methods to improve benchmark quality across tasks and datasets. The researcher will share findings with benchmark authors, collaborate on repairs, and help ensure improved benchmark versions are adopted. Snorkel AI’s Open Benchmarks Grants program is a $30 million commitment to fund and support researchers and independent builders developing AI evaluations. It supports teams behind Terminal-Bench, OSWorld 2.0, and Senior SWE-bench, as well as a Research Fellows program for PhD students and postdocs. Funded work covers areas such as physical AI, scientific discovery, cybersecurity, safety, and alignment, and includes testing and strengthening benchmarks before and after public release. Snorkel AI describes itself as a frontier AI data lab that develops datasets, benchmarks, evaluations, and custom solutions for real-world AI systems. The company was founded out of the Stanford AI Lab in 2019 and works with AI labs and enterprises. It presents the role as an opportunity to contribute to technical research and engineering while influencing priorities and strategic decisions in a growing company.

Requirements

  • Strong research or engineering background in machine learning, LLMs, AI agents, and evaluation.
  • Strong Python programming skills and experience building, running, and debugging AI agents or evaluation systems.
  • Ability to identify weaknesses in existing evaluation methods, develop new approaches, and rigorously test whether they work.
  • Experience designing rigorous experiments, analyzing model behavior, and investigating unexpected failures.
  • Familiarity with LLM or agent benchmarks, evaluation frameworks, or automated grading systems.
  • Experience independently driving research and engineering projects from problem identification through implementation and validation.
  • Strong communication and collaboration skills.

Nice To Haves

  • Experience with adversarial testing, red teaming, reward hacking, or specification gaming.
  • Experience developing or evaluating benchmarks for coding agents, computer use, professional workflows, or similar areas.
  • Familiarity with RL environments, agent trajectories, verifiers, and evaluation infrastructure.
  • Open-source contributions or publications in AI evaluation, robustness, or agentic systems.
  • PhD in Computer Science, Machine Learning, or a related field, or equivalent research or industry experience.

Responsibilities

  • Identify and reproduce benchmark issues such as reward hacking, grader failures, task ambiguity, shortcuts, and environment vulnerabilities.
  • Propose next steps for validated issues, implement selected fixes with benchmark authors, and verify fixes without breaking valid solutions.
  • Reuse existing tools where possible and develop new checks to detect recurring benchmark weaknesses.
  • Review findings with benchmark authors before publication, collaborate on repairs, and support official improved releases.
  • Contribute to technical reports, open-source tools, publications, and new approaches to benchmark evaluation and improvement.

Benefits

  • Meaningful opportunities to shape priorities and initiatives and influence strategic decisions.
  • Support for career growth, including developing technical expertise, exploring leadership opportunities, and learning skills across multiple functions.
  • A company environment described as combining stability with the excitement of high growth.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service