Latest AI JobsBrowse Jobs
Anthropic logo

Research Engineer, Interpretability

AnthropicSan Francisco, CAAI Research & Engineering
Apply on company site
Salary
Not listed
Job type
Full-time
Posted
7 Nov 2025
Status
Open · checked 7d ago

About this role

The Research Engineer, Interpretability position at Anthropic in San Francisco is an engineering role focused on building specialized infrastructure to reverse-engineer neural networks and advance AI safety. This opportunity suits experienced software professionals eager to tackle complex scaling bottlenecks and design tools that empower researchers to decode modern language models.

Role Overview

Anthropic seeks an experienced professional to join the Interpretability team in San Francisco, California. This role centers on reverse-engineering how trained language models work by building sophisticated software tools and infrastructure. The position involves working across the full software stack, from model internals and accelerator-level optimization to user-facing tooling that helps researchers run experiments without hitting engineering barriers. The work directly contributes to safety audits on frontier models and addresses one of the most important open challenges in artificial intelligence.

Key Responsibilities

Build and maintain specialized inference and training infrastructure designed for interpretability research, including instrumented forward and backward passes, activation extraction, and steering vector application.
Resolve scaling and efficiency bottlenecks through profiling, optimization, and close collaboration with peer infrastructure teams.
Design abstractions, platforms, and tooling that enable researchers to rapidly execute experiments without encountering engineering obstacles.
Help integrate interpretability research into production safety audits with high reliability expectations and concrete deadlines.
Work across the entire stack, spanning model internals, user-facing tooling, and accelerator-level optimization.

Requirements

Possess 5 to 10+ years of professional experience building software.
High proficiency in at least one programming language, such as Python, Rust, Go, or Java, alongside strong productivity using Python.
Extreme curiosity regarding unfamiliar domains, with the ability to quickly learn and apply knowledge across new layers of the stack.
Strong capacity to prioritize high-impact work while operating comfortably amidst ambiguity and questioning existing assumptions.
Preference for fast-moving, collaborative projects over extensive solo efforts.
Deep curiosity about interpretability research and its pivotal role in AI safety, with a strong care for societal impacts and ethics.
Ability to work closely with researchers to translate complex research needs into practical engineering solutions.
Minimum education requirement of a Bachelor’s degree in a relevant field of study, or an equivalent combination of education, training, and experience.

Nice to Have

Experience optimizing the performance of large-scale distributed systems.
Knowledge of language modeling fundamentals featuring transformers.
Expertise in High Performance LLM optimization, including memory management, compute efficiency, parallelism strategies, and inference throughput optimization.
Hands-on experience working within a mainstream machine learning stack such as PyTorch and CUDA on GPUs, or JAX and XLA on TPUs.
Background collaborating closely with researchers or performing direct research characterized by complex engineering challenges.

Skills and Tools

Python
PyTorch
Java
Rust
Go
CUDA
JAX
XLA
Large Language Models (LLMs)

Salary and Benefits

The annual compensation range for this role is $315,000 to $560,000 USD. Anthropic provides competitive compensation and benefits, optional equity donation matching, generous vacation and parental leave, flexible working hours, and an office space in San Francisco designed for team collaboration. Visa sponsorship is available for qualified candidates.

About the Company

Anthropic is a public benefit corporation headquartered in San Francisco, California, dedicated to creating reliable, interpretable, and steerable AI systems. The organization views AI research as an empirical science akin to physics and biology, operating as a single cohesive team focused on large-scale research efforts to ensure artificial intelligence remains safe and beneficial for society.

How to Apply

Interested candidates should visit the official Anthropic careers page to submit an application for this position.

Skills and tech stack

PythonPyTorchJavaRustLLM

Questions candidates ask

Where is this role located?

This role is primarily based in the Anthropic office in San Francisco, CA, though exceptional candidates may be considered for remote work on a case-by-case basis.

What is the salary range for the Research Engineer position?

The annual compensation range for this role is between $315,000 and $560,000 USD.

Is prior interpretability research experience required?

No, prior experience working specifically on interpretability research is not required, though curiosity about the field and its role in AI safety is essential.

Does Anthropic offer visa sponsorship for this role?

Yes, Anthropic sponsors visas for eligible candidates and retains an immigration lawyer to assist through the process.

What programming languages are required?

Candidates must be highly proficient in at least one programming language such as Python, Rust, Go, or Java, and be productive with Python.

Ready to apply?

The application is on the official Anthropic careers page.

Apply on company site