Research Engineer, Takeoff Intel

Location: Remote-friendly, San Francisco, United States; staff are expected to work from one of Anthropic's offices at least 25% of the time, with travel required.

Annual salary: $350,000–$850,000 USD

Company

Anthropic develops reliable, interpretable, and steerable AI systems focused on safety and societal benefit.

What you will do

• Design, build, and run capability evaluations and measurement instruments at scale.
• Develop data and analysis pipelines that convert model outputs and telemetry into reliable metrics.
• Prototype, validate, and refine new research instruments quickly.
• Review and supervise AI-written code as part of the engineering workflow.
• Collaborate with research scientists and partner teams to define meaningful measurements.
• Contribute to internal research write-ups and public reporting.

Requirements

• Bachelor's degree or an equivalent combination of education, training, and experience in a relevant field.
• Experience shipping an evaluation, data product, or research library end to end.
• Experience running experiments on large language models.
• Ability to prototype quickly, work with messy large-volume data, and operate from ambiguous research questions.
• Clear communication and close collaboration with research teams.
• Availability to work from an Anthropic office at least 25% of the time and travel as required.

Nice to have

• Experience building LLM evaluation harnesses or benchmark infrastructure.
• Large-scale ML or data infrastructure experience.
• Experience building tools or libraries used by other researchers.
• Experience identifying errors in AI-written code.

Culture & Benefits

• Collaborative research environment focused on high-impact AI safety and empirical science.
• Frequent research discussions and close collaboration across research, policy, economics, pretraining, and reinforcement learning teams.
• Flexible working hours and a collaborative office environment.
• Competitive compensation, equity donation matching, generous vacation, and parental leave.
• Visa sponsorship is available, subject to role and candidate eligibility.

Hiring process

• Applications are evaluated with consideration for equivalent education, training, and experience.
• Candidates may use AI in the application process according to Anthropic's candidate AI usage guidance.