ML Performance Engineer: Scale & Optimize GPU/CPU Workloads

G-Research; Greater London; On-site; GBP 90,000 - 135,000; Full time

Benefits offered by this job

Lunch provided (Just Eat for Business)

Barista bar

35 days annual leave

9% pension contributions

Informal dress code

Healthcare and life assurance

Cycle-to-work scheme

Monthly company events

Job summary

G-Research is seeking an exceptional ML Performance Engineer in London to optimize large-scale workloads across GPU and CPU infrastructure. You will profile, benchmark and tune training and inference workloads, design reference implementations, and collaborate with researchers and infrastructure teams to evolve the compute stack.

You will contribute to a scalable platform supporting cutting-edge machine learning research, with strong emphasis on performance, reliability and cross-team

Qualifications

• Bachelor's/Master's/PhD in CS or equivalent experience.
• Proven track record in profiling, benchmarking and optimizing distributed workloads.
• Experience with Python and CUDA.
• Experience with HPC schedulers and Kubernetes-based orchestration.
• Strong knowledge of PyTorch or similar DL frameworks.
• Strong understanding of Linux fundamentals (scheduling, memory, NUMA, filesystems).

Responsibilities

• Collaborate with researchers and engineers to design optimized solutions for compute challenges.
• Profile, benchmark and tune large-scale training and inference workloads across CPU/GPU/memory.
• Develop reference implementations, libraries and tools to improve job efficiency and reliability.
• Collaborate with systems and platform teams to evolve the compute stack.
• Influence long-term platform and infrastructure decisions.

Skills

Python

Deep learning frameworks

Linux systems

Data structures & algorithms

Parallel programming

Communication skills

Education

Bachelor's/Master's/PhD in CS or equivalent

Tools

CUDA

nsys / ncu

eBPF tools

Kubernetes

HPC schedulers

Job description

G-Research is seeking an exceptional ML Performance Engineer in London to optimize large-scale workloads across GPU and CPU infrastructure. You will profile, benchmark and tune training and inference workloads, design reference implementations, and collaborate with researchers and infrastructure teams to evolve the compute stack.

You will contribute to a scalable platform supporting cutting-edge machine learning research, with strong emphasis on performance, reliability and cross-team