Research Engineer, Interpretability
$315k - $560k • Remote • San Francisco, CA
Posted 9d ago
About the job
The Interpretability team at Anthropic is dedicated to understanding how modern language models work, believing that a deep, mechanistic understanding is crucial for making advanced AI systems safe and steerable. This role involves building and maintaining the specialized infrastructure that powers interpretability research, akin to performing 'neuroscience' on neural networks. The work spans the entire lifecycle of a production language model, from pretraining and inference to performance optimization, and directly contributes to safety audits on frontier models.
Responsibilities
- Build and maintain specialized inference and training infrastructure for interpretability research, including instrumented forward/backward passes, activation extraction, and steering vector application.
- Resolve scaling and efficiency bottlenecks through profiling, optimization, and collaboration with peer infrastructure teams.
- Design tools, abstractions, and platforms to enable rapid researcher experimentation.
- Integrate interpretability research into production safety audits with high reliability expectations.
- Work across the full stack, from model internals and accelerator-level optimization to user-facing research tooling.
Requirements
- 5-10+ years of experience building software.
- Highly proficient in at least one programming language (e.g., Python, Rust, Go, Java) and productive with Python.
- Extremely curious about unfamiliar domains and able to quickly learn and apply new knowledge.
- Strong ability to prioritize impactful work and comfortable operating with ambiguity.
- Prefers fast-moving collaborative projects.
- Curious about interpretability research and its role in AI safety.
- Cares about the societal impacts and ethics of work.
- Comfortable working closely with researchers and translating research needs into engineering solutions.