Software Engineer - Data & Evaluation, Horizons

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 Horizons

The Horizons team leads Anthropic's reinforcement learning research and development, playing a critical role in advancing our AI systems. We've contributed to all Claude models, with significant impacts on the autonomy and coding capabilities of Claude 3.5 and 3.7 Sonnet. Our work spans several key areas:

• Developing systems that enable models to use computers effectively
• Advancing code generation through reinforcement learning
• Pioneering fundamental RL research for large language models
• Building scalable RL infrastructure and training methodologies
• Enhancing model reasoning capabilities

We collaborate closely with Anthropic's alignment and frontier red teams to ensure our systems are both capable and safe. We partner with the applied production training team to bring research innovations into deployed models, and work hand-in-hand with dedicated RL engineering teams to implement our research at scale. The Horizons team sits at the intersection of cutting-edge research and engineering excellence, with a deep commitment to building high-quality, scalable systems that push the boundaries of what AI can accomplish.

About the Role

As a Data & Evaluation Engineer on the Horizons team, you will build the software infrastructure that enables our AI models to use tools effectively and measure their performance. You'll develop and extend our agent framework, create and implement evaluations, manage training data pipelines, and apply data science techniques to improve model capabilities. This engineering-focused role combines software development with empirical analysis to drive advances in model performance and capabilities.

Representative projects:

• Extend and improve our agent framework that enables models to interact with tools and environments
• Design and implement evaluation systems that rigorously measure model capabilities across tasks
• Build and maintain data pipelines for collecting, processing, and managing RL training data
• Develop dashboards and analysis tools to extract insights from model performance data
• Collaborate with researchers to translate evaluation needs into scalable, production-grade systems

You may be a good fit if you:

• Are proficient in Python and data analysis libraries (Pandas, NumPy, etc.)
• Have experience creating data visualizations and interactive dashboards
• Enjoy translating complex research questions into concrete measurement approaches
• Possess strong software engineering fundamentals with an emphasis on clean APIs
• Can communicate technical concepts clearly to both engineers and researchers
• Have experience with web development for interactive tools (JavaScript, React, etc.)
• Are passionate about measuring and improving AI capabilities and safety

Strong candidates may have:

• Experience with LLM specific evaluations and frameworks
• Experience with data visualization libraries and frameworks (D3.js, Plotly, Grafana)
• Familiarity with the Jupyter ecosystem and notebook-based workflows
• Knowledge of statistical methods and experimental design
• Experience with web frameworks for building interactive applications (FastAPI, Flask)
• Experience working with large datasets and understand performance considerations
• Comfort engaging with ML research papers and implementing metrics from academic literature

Strong candidates need not have:

• Formal certifications or education credentials
• Prior experience with reinforcement learning research

Deadline to apply: None. Applications will be reviewed on a rolling basis.

The expected salary range for this position is:

Annual Salary:$300,000—$425,000 USD

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.