Software Engineer, ML Performance and Scaling: Anthropic

Anthropic, a leading AI safety and research company with the mission to create reliable, interpretable, and steerable AI systems, is hiring a Software Engineer for its ML Performance and Scaling team. This team builds the infrastructure that powers Anthropic's largest-scale pre-training runs for its AI models.

As a Performance Engineer, you will be responsible for identifying and solving novel systems problems to optimize the throughput and robustness of the company's largest distributed systems. This role is at the intersection of research, performance, and distributed systems, and is ideal for a strong software engineer who is excited to become an expert in large-scale machine learning systems.

Key Responsibilities:
Your work will involve projects such as:

Implementing low-latency, high-throughput sampling for large language models.

Implementing GPU kernels to adapt models for low-precision inference.

Writing custom load-balancing algorithms to optimize serving efficiency.

Building quantitative models of system performance.

Designing and implementing fault-tolerant distributed systems with complex network topologies.

Debugging kernel-level network latency spikes in containerized environments.

Qualifications & Experience:
Required:

Significant software engineering or machine learning experience, particularly at supercomputing scale.

A results-oriented mindset with a bias towards flexibility and impact.

An enjoyment of pair programming and a collaborative spirit.

A genuine care for the societal impacts of your work.

Preferred:

Experience with high-performance, large-scale ML systems.

GPU/Accelerator programming.

Deep knowledge of ML framework internals or OS internals.

Experience with language modeling with transformers.

Note: Anthropic sponsors visas for this role.

Why Join Anthropic?
Work as part of a single cohesive team on a few large-scale research efforts with high impact.

Be part of an extremely collaborative group focused on advancing steerable and trustworthy AI.

The company offers competitive compensation, generous vacation and parental leave, and flexible working hours.