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 are the team behind the Responsible Scaling Policy (RSP) Evaluations. The easiest way to understand what we do is to read the RSP sections of the Claude 4 system card. We build and operate the automated systems that test whether frontier AI models are safe to release. Our evaluations determine if models have crossed critical capability thresholds in domains like autonomous replication, cybersecurity, and biological research. We're a team of builders who switch into high-stakes execution mode during model launches—we don't just build the evaluations, we run them.
This is a production engineering role focused on building reliable evaluation infrastructure for AI safety. You'll create automated systems that systematically test frontier AI models for dangerous capabilities, working closely with domain experts to translate safety concerns into robust, scalable evaluation pipelines.
What makes this role unique- High-stakes execution: Your systems directly determine whether cutting-edge AI models get released
- Model launch cadence: You'll build during development cycles and execute flawlessly during critical launch windows
- Production reliability: Build evaluation infrastructure that works consistently under pressure
- Cross-domain collaboration: Work with biosecurity, cybersecurity, and other domain experts to automate their safety assessments
- Build and maintain automated evaluation systems using distributed infrastructure
- Create robust evaluation pipelines that can run thousands of model capability tests
- Develop tools that allow domain experts to quickly deploy new safety evaluations
- Ensure evaluation systems run reliably during high-stakes model launches
- Write production-quality Python code for evaluation infrastructure that scales
- Monitor and operate evaluation systems during critical assessment periods
- Collaborate with domain experts to translate safety requirements into technical implementations
- Are a reliable engineer who builds systems that work consistently under pressure
- Have strong Python programming skills and write clean, maintainable code
- Are comfortable working with LLMs programmatically (APIs, prompting, output processing)
- Have experience with debugging complex systems and resolving issues quickly
- Can work independently on 1-3 month projects while collaborating effectively with domain experts
- Understand systems optimization and can build efficient, scalable solutions
- Care about AI safety and want to contribute to responsible model development
- Thrive in environments that switch between building and execution modes
- Experience with evaluation frameworks, testing infrastructure, or automated assessment systems
- Background in physics, systems engineering, or other fields requiring critical thinking about experimental results
- Familiarity with distributed systems, containerization, or production operations
- Experience with adversarial testing, red-teaming, or finding edge cases in complex systems
- Understanding of LLM capabilities and limitations
- Direct impact: Your systems determine whether the world's most advanced AI models get released
- Technical growth: Build expertise in AI safety evaluation while working with cutting-edge models
- Collaborative environment: Work with world-class domain experts across multiple safety-critical fields
- Production ownership: Own the full lifecycle from building evaluation systems to operating them during launches
Mission-driven work: Contribute directly to ensuring AI systems remain safe and beneficial
Representative Projects- Build automated red-teaming systems that generate and evaluate thousands of adversarial prompts
- Create evaluation pipelines that systematically test model capabilities across multiple risk domains
- Develop monitoring infrastructure that tracks evaluation results and detects capability jumps
- Implement reliable containerized environments for running large-scale model assessments
- Build tools that allow biosecurity experts to quickly create and deploy new biological risk evaluations
- Create automated analysis systems that process evaluation results and generate capability reports
- Previous AI/ML research experience
- Domain expertise in specific risk areas like biosecurity or cybersecurity
- Senior-level experience - we're looking for someone ready to grow into owning critical infrastructure
Deadline to apply: None. Applications will be reviewed on a rolling basis.
The expected salary range for this position is:
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 differentWe 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
Similar Jobs
What We Do
Anthropic is an AI safety and research company that’s working to build reliable, interpretable, and steerable AI systems. Our research interests span multiple areas including natural language, human feedback, scaling laws, reinforcement learning, code generation, and interpretability.
