Research Engineer – Interpretability

About the role

Anthropic is looking for a Research Engineer to join its Interpretability team. You will help reverse‑engineer large language models to build a mechanistic understanding that underpins safe and trustworthy AI. The work blends cutting‑edge research with production‑grade engineering to scale interpretability tools for real‑world safety audits.

Key responsibilities

• Build and maintain specialized inference and training infrastructure for interpretability research, including instrumented forward/backward passes, activation extraction, and steering‑vector application.
• Identify and resolve scaling and efficiency bottlenecks through profiling, optimization, and close collaboration with peer infrastructure teams.
• Design abstractions, tools, and platforms that enable researchers to experiment rapidly without hitting engineering barriers.
• Help transition interpretability research findings into production safety audits and related pipelines.

Required profile

• Experience building and operating large‑scale model training and inference infrastructure.
• Strong ability to profile systems, diagnose performance issues, and implement optimizations.
• Collaborative mindset with a track record of working closely with infrastructure and research teams.
• Interest in AI safety and interpretability research.

Required skills