Research Engineer, Model Evaluations

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

As a Research Engineer on the Model Evaluations team, you'll lead the design and implementation of Anthropic's evaluation platform—a critical system that shapes how we understand, measure, and improve our models' capabilities and safety. You'll work at the intersection of research and engineering to develop and implement model evaluations that give us insight into emerging capabilities and build robust evaluation infrastructure that directly influences our training decisions and model development roadmap.

Your work will be essential to Anthropic's mission of building safe, beneficial AI systems. You'll collaborate closely with training teams, alignment researchers, and safety teams to ensure our models meet the highest standards before deployment. This is a technical leadership role where you'll drive both the strategic vision and hands-on implementation of our evaluation systems.

Responsibilities

• Design novel evaluation methodologies to assess model capabilities across diverse domains including reasoning, safety, helpfulness, and harmlessness
• Lead the design and architecture of Anthropic's evaluation platform, ensuring it scales with our rapidly evolving model capabilities and research needs
• Implement and maintain high-throughput evaluation pipelines that run during production training, providing real-time insights to guide training decisions
• Analyze evaluation results to identify patterns, failure modes, and opportunities for model improvement, translating complex findings into actionable insights
• Partner with research teams to develop domain-specific evaluations that probe for emerging capabilities and potential risks
• Build infrastructure to enable rapid iteration on evaluation design, supporting both automated and human-in-the-loop assessment approaches
• Establish best practices and standards for evaluation development across the organization
• Mentor team members and contribute to the growth of evaluation expertise at Anthropic
• Coordinate evaluation efforts during critical training runs, ensuring comprehensive coverage and timely results
• Contribute to research publications and external communications about evaluation methodologies and findings

You may be a good fit if you

• Have experience designing and implementing evaluation systems for machine learning models, particularly large language models
• Have demonstrated technical leadership experience, either formally or through leading complex technical projects
• Are skilled at both systems engineering and experimental design, comfortable building infrastructure while maintaining scientific rigor
• Have strong programming skills in Python and experience with distributed computing frameworks
• Can translate between research needs and engineering constraints, finding pragmatic solutions to complex problems
• Are results-oriented and thrive in fast-paced environments where priorities can shift based on research findings
• Enjoy collaborative work and can effectively communicate technical concepts to diverse stakeholders
• Care deeply about AI safety and the societal impacts of the systems we build
• Have experience with statistical analysis and can draw meaningful conclusions from large-scale experimental data

Strong candidates may also have

• Experience with evaluation during model training, particularly in production environments
• Familiarity with safety evaluation frameworks and red teaming methodologies
• Background in psychometrics, experimental psychology, or other fields focused on measurement and assessment
• Experience with reinforcement learning evaluation or multi-agent systems
• Contributions to open-source evaluation benchmarks or frameworks
• Knowledge of prompt engineering and its role in evaluation design
• Experience managing evaluation infrastructure at scale (thousands of experiments)
• Published research in machine learning evaluation, benchmarking, or related areas

Representative projects

• Designing comprehensive evaluation suites that assess models across hundreds of capability dimensions
• Building real-time evaluation dashboards that surface critical insights during multi-week training runs
• Developing novel evaluation approaches for emerging capabilities like multi-step reasoning or tool use
• Creating automated systems to detect regression in model performance or safety properties
• Implementing efficient evaluation sampling strategies that balance coverage with computational constraints
• Collaborating with external partners to develop industry-standard evaluation benchmarks
• Building infrastructure to support human evaluation at scale, including quality control and aggregation systems 

Logistics

Education requirements: We require at least a Bachelor's degree in a related field or equivalent experience.

Location-based hybrid policy: Currently, we expect all staff to be in one of our offices at least 25% of the time. However, some roles may require more time in our offices.

Visa sponsorship: We do sponsor visas! However, we aren't able to successfully sponsor visas for every role and every candidate. But if we make you an offer, we will make every reasonable effort to get you a visa, and we retain an immigration lawyer to help with this.

We encourage you to apply even if you do not believe you meet every single qualification. Not all strong candidates will meet every single qualification as listed.  Research shows that people who identify as being from underrepresented groups are more prone to experiencing imposter syndrome and doubting the strength of their candidacy, so we urge you not to exclude yourself prematurely and to submit an application if you're interested in this work. We think AI systems like the ones we're building have enormous social and ethical implications. We think this makes representation even more important, and we strive to include a range of diverse perspectives on our team.

How we're different

We believe that the highest-impact AI research will be big science. At Anthropic we work as a single cohesive team on just a few large-scale research efforts. And we value impact — advancing our long-term goals of steerable, trustworthy AI — rather than work on smaller and more specific puzzles. We view AI research as an empirical science, which has as much in common with physics and biology as with traditional efforts in computer science. We're an extremely collaborative group, and we host frequent research discussions to ensure that we are pursuing the highest-impact work at any given time. As such, we greatly value communication skills.

The easiest way to understand our research directions is to read our recent research. This research continues many of the directions our team worked on prior to Anthropic, including: GPT-3, Circuit-Based Interpretability, Multimodal Neurons, Scaling Laws, AI & Compute, Concrete Problems in AI Safety, and Learning from Human Preferences.

Come work with us!

Anthropic is a public benefit corporation headquartered in San Francisco. We offer competitive compensation and benefits, optional equity donation matching, generous vacation and parental leave, flexible working hours, and a lovely office space in which to collaborate with colleagues. Guidance on Candidates' AI Usage: Learn about our policy for using AI in our application process