Research Scientist, Takeoff Intel

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.

We're looking for a Research Scientist who has done hands-on research on large models (pretraining, fine-tuning, RL, evals, or agents scaffolds) and wants to focus on measuring and understanding recursive-self-improvement. You know what the model-development loop looks like from the inside: which signals matter and where the real bottlenecks are. On this team you'll use that judgment to decide what's worth measuring, design the evaluations and models that measure it, and interpret what the results mean for how fast this is moving.

We're hiring at both junior and senior levels. Senior researchers should be comfortable doing hands-on technical work alongside setting research direction.

• Identify the signals that track AI R&D acceleration and design the evaluations that measure them

• Build quantitative models of capability growth and self-improvement dynamics, grounded in evaluation and telemetry data

• Run experiments and evals to test hypotheses about automation and capability

• Make opinionated research bets and own the outcome

• Write graded assessments of what our measurements show, for internal decision-makers and public reporting

• Collaborate with pretraining, RL, economic research, and policy teams

You may be a good fit if you

• Have done hands-on research on large language models: pretraining, fine-tuning, RL, evals, or agent systems

• Have strong quantitative instincts, are comfortable with quantitative modeling and reasoning

• Have experience in forecasting, may have published AI forecasting scenarios

• Can design an evaluation from a vague question and defend the methodology

• Write clearly and calibrate: state confidence, name what would change your conclusion

• Are motivated by impact: comfortable with work whose output is graded assessments and system-card sections more often than papers

• Care about AI safety and think carefully about where rapid capability growth leads

Strong candidates may also have

• Trained or RL'd frontier models hands-on

• Experience with scaling laws, capability forecasting, or emergent-capability studies

• A physics, applied-math, or similarly quantitative background that moved into ML

• Written a system card section, capability report, or methodology document that others cite

• Experience supervising and correcting AI-written code

Some examples of our work

• Anthropic ECI: our adaptation of Epoch Capabilities Index published in all recent system cards to measure capability acceleration

• AI R&D capability assessments in the Claude system cards

• When AI Builds Itself: all data in the article comes from our team

The annual compensation range for this role is listed below.

For sales roles, the range provided is the role’s On Target Earnings ("OTE") range, meaning that the range includes both the sales commissions/sales bonuses target and annual base salary for the role.

Annual