Research Engineer, Rule of Law

Details

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 Anthropic Institute

Anthropic is a public benefit corporation building some of the world's most powerful AI systems. The Anthropic Institute is a new externally-facing organization within Anthropic whose purpose is to give the world new information about how the AI systems we build are affecting the economy, democratic institutions, and the people and organizations that interact with our systems. The Institute sits inside a frontier lab, with access to information only AI developers possess. That position is what makes our work different. We don't just study AI from the outside. We study it from within.

About the role

As a Research Engineer on the Rule of Law team, you'll conduct technical and sociotechnical research at the intersection of AI and democratic institutions. Your work will span safety evaluations, model improvement, institutional analysis, and the development of novel applications of AI to support civic life and efficient and accountable government. The work will directly contribute to research publications, policy work, safety systems, and products.

Current focus areas include:

• Assuring that AI agents rigorously obey the law: evaluations and fine-tuning to assess and improve how models process legal constraints, including alignment work in settings where defining the optimization objective is hard and novel techniques may be needed.
• Helping to ensure that AI reinforces the foundations of constitutional democracy: understanding how AI will impact the structure of government itself, and developing approaches to uphold popular sovereignty, ethical and accountable government, and the rule of law.
• Enriching and expanding civic participation: using AI to support democratic deliberation, improve citizen–government interfaces, and help citizens deliberate with one another in an informed and constructive way around policy issues.

Responsibilities

• Design and run AI safety evaluations focused on legal alignment, and conduct technical work (including fine-tuning) to improve targeted aspects of model performance.
• Leverage AI to analyze institutional vulnerabilities created or exacerbated by AI itself, and develop accompanying sociotechnical mitigations.
• Develop novel applications of AI to bolster and enrich democratic processes.
• Conduct rigorous sociotechnical studies on how information furnished by AI can support informed deliberation and debate.
• Partner closely with researchers, policy and law experts, and other cross-functional teams across Anthropic to advance the safety mission.
• Translate research insights into actionable recommendations for both product and policy.

Requirements / Qualifications

• Deep expertise in AI together with substantive expertise in government, law, political science, or public policy.
• Advanced skills in deep learning, together with capabilities in data science, mechanism design, govtech, deliberative tech, or a related field.
• Deep understanding of experimental design, data analysis, and inferential statistics.
• At least five years of work experience in academia, industry, or government.
• Comfortable navigating ambiguity inherent to novel research, including settings where the objective itself is hard to define.
• Strong technical skills in sociotechnical settings: thinking carefully about human factors, how AI systems interact with groups of humans, and human behavior.
• Energized by working at the intersection of technical research and questions of democracy and institutional design.

Strong candidates may also have

• Experience designing and running evaluations or fine-tuning experiments on large language models.
• Large-scale data analysis experience.
• UX and interface design experience, with an eye toward making complex tools intuitive for non-technical users.
• Engineering experience with privacy considerations front of mind.
• Research experience relating to deliberative technology, civic technology, or the societal impacts of AI.
• Experience working with or within government, courts, legislatures, or public policy institutions.

Representative projects

• AI safety evaluations focusing on legal alignment, with technical work to improve targeted aspects of model performance.
• Analysis of institutional vulnerabilities, with accompanying work toward technical mitigations.
• Development of novel applications for AI to bolster and enrich democratic processes.
• Sociotechnical work focused on how information furnished by AI can support informed deliberation and debate.

Compensation

Annual Salary: $320,000 - $485,000 USD

Logistics

• Minimum education: Bachelor’s degree or an equivalent combination of education, training, and/or experience.
• Required field of study: A field relevant to the role as demonstrated through coursework, training, or professional experience.
• Minimum years of experience: Years of experience required will correlate with the internal job level requirements for the position (posting indicates at least five years of experience).
• Location-based hybrid policy: Currently, staff are expected to be in one of our offices at least 25% of the time; some roles may require more time in offices.
• Visa sponsorship: Anthropic states they sponsor visas and retain an immigration lawyer to help with this, though not every role/candidate can be successfully sponsored.

How we're different

Anthropic emphasizes big-science AI research, collaborative work on a few large-scale research efforts, and impact-driven research toward steerable, trustworthy AI. Frequent research discussions and strong communication skills are valued.

Come work with us / Benefits

Anthropic is headquartered in San Francisco and offers competitive compensation and benefits, optional equity donation matching, generous vacation and parental leave, flexible working hours, and a collaborative office space. Guidance on candidates' AI usage is provided in the application process.