About this role
About Anthropic
Anthropic’s mission is to create reliable, interpretable, and steerable AI systems. We want AI to be safe and beneficial for our users and for society as a whole. Our team is a quickly growing group of committed researchers, engineers, policy experts, and business leaders working together to build beneficial AI systems.
About the role
Anthropic's Life Sciences team is building a research group focused on biological discoveries by combining AI with hands-on biological research. We're seeking a research scientist who combines deep chemistry knowledge with machine learning expertise to join the team. Our goal is to speed up progress in the life sciences by an order of magnitude, and this role puts that mission into practice. You'll be a computational chemist who knows reactions, synthesis, and medicinal chemistry in depth and has hands-on experience training and evaluating machine learning models. Above all, you'll build the reinforcement learning environments, datasets, and evaluations that teach Claude chemistry. Working alongside our bench chemists and AI researchers, you'll turn real chemistry problems and experimental data into training tasks and measure whether the models get better at them.
This role is fully on-site in the San Francisco Bay Area.
Key responsibilities
• Design and build reinforcement learning environments for chemistry end to end, including task design, data curation, automated grading, and checks against shortcuts and data leakage
• Design and source new experimental datasets, environments, and evaluations with our bench chemists, and external research and academic partners
• Develop evaluations that measure model performance on real chemistry tasks, such as reaction outcome prediction, synthesis planning, structure elucidation, and compound design, and feed results back to research teams
• Run and analyze training experiments with AI research teams, diagnose where models fail at chemistry, and prioritize what to build next
• Ensure datasets and environments are high quality and well documented, including sources, train and test splits, and failure modes
• Communicate findings clearly to chemistry, AI research, and life sciences audiences
Minimum qualifications
• PhD in chemistry, computational chemistry, or a related field
• Deep knowledge of organic and medicinal chemistry, including reactions, synthesis, and SAR, with experimental data as your ground truth
• Strong Python skills and solid machine learning fundamentals, with hands-on experience training and evaluating models
• Comfort navigating ambiguity and defining problems in a rapidly evolving research environment
• Ability to work independently while collaborating effectively across cross-functional teams, with strong written and verbal communication skills
Preferred qualifications
• Experience with reinforcement learning, large language models, or building evaluations and benchmarks
• Track record of supporting drug discovery pro