Staff+ Software Engineer, Safeguards Evals

Anthropic•Seattle, WA

1d•$320,000 - $485,000•Hybrid

About The Position

Anthropic is hiring a Staff+ Software Engineer to develop methods and infrastructure for trustworthy evaluations of model behavior and agentic systems. The role sits at the intersection of applied machine learning research and engineering, designing experiments to improve evaluation quality, building datasets grounded in real-world misuse, and deploying evaluation methods into pipelines that inform model training, agent changes, and launch decisions. The role also owns evaluation of an agentic investigation system, including its metrics, test cases, grading approaches, and end-to-end performance. The engineer will build tools and pipelines that allow policy experts to author and run evaluations, and will share findings with research and training teams to support model improvements. The work covers areas such as cyber attacks, biological weapons, and influence operations, and includes collaboration with Policy and Enforcement.

Requirements

Experience building and maintaining data pipelines.
Experience working with LLMs and a working understanding of their capabilities and failure modes, especially agentic systems with tool use and multi-step reasoning.
Strong data analysis skills and the ability to draw reliable insights from large datasets.
Ability to move between research prototyping and production-quality code.
Ability to translate ambiguous problems into concrete, testable experiments.
Care deeply about AI safety and want your work to have real impact.
At least a bachelor's degree or an equivalent combination of education, training, and/or experience.
A field of study relevant to the role, demonstrated through coursework, training, or professional experience.

Nice To Haves

8 or more years of industry software engineering or machine learning engineering experience.
Expertise in building or contributing to LLM or agent evaluation frameworks, benchmarks, or automated grading systems.
Extensive experience in trust and safety, content moderation, or abuse detection systems.
Experience in red teaming, adversarial testing, or jailbreak research on AI systems.
Experience with synthetic data generation or data augmentation.
Experience with distributed systems or large-scale data processing.
Experience with prompt engineering or building LLM-powered applications.

Responsibilities

Design and run experiments to improve evaluation quality, including methods for generating representative test data, simulating realistic user behavior, and validating grading accuracy.
Research how multi-turn conversations, tools, long context, and user diversity affect model safety behavior.
Analyze evaluation coverage to identify measurement gaps and keep evaluations unsaturated and high-signal as model and agent capabilities advance.
Build and own the evaluation harness for an agentic investigation system, defining metrics, test cases, and grading approaches.
Measure agent performance end-to-end, including detection precision and recall, investigation quality, and robustness, and improve performance on difficult harm areas.
Construct reinforcement-learning environments to improve Claude's safety investigation capabilities.
Build high-quality evaluation datasets representing real-world misuse, drawing on real traffic patterns and synthetic generation.
Collaborate with Policy and Enforcement to turn observed harm patterns into measurable evaluations.
Deploy successful research into evaluation, regression, and release pipelines that run during model training, after agent changes, prompt updates, and underlying model upgrades, as well as beyond launch.
Build tools that enable policy experts to author, run, and iterate on evaluations without engineering support.
Surface evaluation findings to research and training teams to drive upstream model improvements.