Research Scientist, Takeoff Intel

Details

Anthropic is seeking a Research Scientist with hands-on experience researching large models, including pretraining, fine-tuning, reinforcement learning, evaluations, or agent scaffolds. The role focuses on measuring and understanding recursive self-improvement by designing evaluations and models, interpreting results, and assessing the pace of AI research and development progress.

The team is hiring at both junior and senior levels. Senior researchers should be comfortable performing hands-on technical work while setting research direction.

Responsibilities

• Identify signals that track AI R&D acceleration and design evaluations to measure them.
• Build quantitative models of capability growth and self-improvement dynamics using evaluation and telemetry data.
• Run experiments and evaluations to test hypotheses about automation and capability.
• Make opinionated research bets and own the outcomes.
• Write graded assessments of measurement results for internal decision-makers and public reporting.
• Collaborate with pretraining, reinforcement learning, economic research, and policy teams.

Requirements

• Hands-on research experience with large language models, including pretraining, fine-tuning, reinforcement learning, evaluations, or agent systems.
• Strong quantitative instincts and comfort with quantitative modeling and reasoning.
• Experience in forecasting; experience publishing AI forecasting scenarios may be beneficial.
• Ability to design an evaluation from a vague question and defend its methodology.
• Clear writing and calibration, including the ability to state confidence and identify what would change a conclusion.
• Motivation to work on impact-focused research whose outputs may be graded assessments and system-card sections rather than papers.
• Careful consideration of AI safety and the implications of rapid capability growth.

Strong candidates may also have:

• Hands-on experience training or reinforcement-learning frontier models.
• Experience with scaling laws, capability forecasting, or emergent-capability studies.
• A background in physics, applied mathematics, or a similarly quantitative field that transitioned into machine learning.
• Experience writing a system-card section, capability report, or methodology document cited by others.
• Experience supervising and correcting AI-written code.

Minimum education is a bachelor's degree or an equivalent combination of education, training, and experience. The required field of study is a field relevant to the role, as demonstrated through coursework, training, or professional experience. Required years of experience correlate with the internal job-level requirements.

Compensation

The annual salary range is $350,000–$850,000 USD.

Work Policy and Logistics

Anthropic currently expects all staff to work from one of its offices at least 25% of the time, although some roles may require more office time.

Anthropic explicitly sponsors visas for this role where possible and will make every reasonable effort to obtain a visa for candidates who receive an offer. The company retains an immigration lawyer to assist with visa sponsorship.

About Anthropic

Anthropic's mission is to create reliable, interpretable, and steerable AI systems that are safe and beneficial for users and society. Anthropic is a public benefit corporation headquartered in San Francisco. The company offers competitive compensation and benefits, optional equity donation matching, generous vacation and parental leave, flexible working hours, and an office space for collaboration.