High-Throughput ML Systems Performance Engineer

Anthropic; York and North Yorkshire; On-site; GBP 110,000 - 150,000; Full time

Benefits offered by this job

Health insurance

Fertility benefits

Parental leave 22 weeks

Flexible PTO

Mental health support

Equity package

Equity donation matching

Retirement plans

Life and income protection

Wellness stipend

Commuter benefits

Education stipend

Home office stipend

Relocation support

In-office meals

Job summary

Anthropic in the United Kingdom is seeking a Performance Engineer to identify novel systems problems and build high-throughput, low-latency distributed services for large language models. You will implement sampling, GPU kernel adaptations for low-precision inference, custom load-balancing, and fault-tolerant architectures, collaborating with teams to maximize throughput and reliability.

This role requires deep systems expertise at supercomputing scale and a readiness to pair-program and own

Qualifications

• Significant software engineering or ML experience, especially at scale.
• Experience solving large-scale systems problems.
• Interest in ML and societal impact, with results-driven mindset.

Responsibilities

• Implement low-latency high-throughput sampling for large language models.
• Implement GPU kernels to adapt models to low-precision inference.
• Write a custom load-balancing algorithm to optimize serving efficiency.
• Build quantitative models of system performance.
• Design and implement fault-tolerant distributed systems with complex networks.
• Debug kernel-level network latency spikes in containerized environments.

Skills

Low-latency systems

GPU kernel programming

Distributed systems

Load balancing

Kernel debugging

Performance modeling

Tools

Docker

Kubernetes

Job description

Anthropic in the United Kingdom is seeking a Performance Engineer to identify novel systems problems and build high-throughput, low-latency distributed services for large language models. You will implement sampling, GPU kernel adaptations for low-precision inference, custom load-balancing, and fault-tolerant architectures, collaborating with teams to maximize throughput and reliability.

This role requires deep systems expertise at supercomputing scale and a readiness to pair-program and own