Anthropic is seeking a Research Scientist to join their Frontier Red Team, focusing on advanced autonomy evaluations. The role involves developing and implementing evaluations to determine the AI Safety Level (ASL) of Anthropic's models. This is crucial for training, deploying, and securing their models, as outlined in their Responsible Scaling Policy (RSP).
The Research Scientist will lead the end-to-end development of autonomy evaluations, starting with risk and capability modeling, and including designing, implementing, and regularly running these evaluations. They will iterate on experiments to evaluate autonomous capabilities and forecast future capabilities. The role also involves providing technical leadership to Research Engineers to build scalable and secure infrastructure for large-scale experiments.
Anthropic values collaboration and communication, and the Research Scientist will communicate evaluation outcomes to relevant teams, policy stakeholders, and research collaborators. They will also collaborate with other projects to improve infrastructure and design safety techniques for autonomous capabilities.
Responsibilities:
Requirements:
What Anthropic Offers: