Biological Safety Research Scientist

Who you are

You are a biological scientist with a strong understanding of safety mechanisms in AI systems. You have experience in designing and executing capability evaluations to assess the performance of AI models. Your background allows you to translate complex biosecurity concepts into practical technical safeguards. You thrive in collaborative environments, working closely with threat modeling experts and machine learning engineers to develop robust safety systems. You are committed to ensuring that AI technologies are safe and beneficial for society. You understand the balance between advancing legitimate life sciences research and preventing misuse by sophisticated threat actors.

Desirable

Experience in AI safety or biosecurity is a plus. Familiarity with machine learning concepts and practices will enhance your ability to contribute effectively to the team. You are open to learning and adapting to new challenges in the rapidly evolving field of AI safety.

What you'll do

In this role, you will design and execute capability evaluations to assess the capabilities of new AI models. You will collaborate closely with internal and external experts to develop training data for safety systems, ensuring they are optimized for robustness against adversarial attacks while maintaining low false-positive rates for legitimate researchers. You will analyze the performance of safety systems and contribute to the development of oversight mechanisms that align with Anthropic's mission of creating reliable and interpretable AI systems. Your work will directly impact how frontier AI models handle dual-use biological knowledge, shaping the future of responsible AI safety.

What we offer

At Anthropic, we provide competitive compensation and benefits, including optional equity donation matching, generous vacation and parental leave, and flexible working hours. You will work in a collaborative office space in San Francisco, surrounded by a team of dedicated researchers and engineers. We encourage you to apply even if your experience doesn't match every requirement, as we value diverse perspectives and backgrounds in our mission to build beneficial AI systems.