Founding GPU Engineer

Fuse Energy; Greater London; On-site; GBP 90,000 - 130,000; Full time

Benefits offered by this job

Equity sign-on bonus

Biannual bonus

Fully expensed tech

Breakfast and dinner allowance

Job summary

Fuse Energy is seeking a seasoned CUDA performance engineer to design and optimise kernels for high-throughput workloads in a high-performance compute environment.

You will profile GPU bottlenecks, build tooling to align power draw with energy pricing, and optimise multi-GPU scaling across NCCL/MPI. Collaboration with ML and systems teams is essential to improve training and inference pipelines.

Qualifications

• 4+ years of experience writing production CUDA code or equivalent.
• Deep understanding of GPU architecture (SMs, warps, memory hierarchy).
• Proficiency in C++ and CUDA; Python for tooling/automation.
• Experience with performance profiling tools (Nsight Systems/Compute).
• Familiarity with multi-GPU/multi-node scaling (NCCL, MPI, RDMA/InfiniBand).
• Strong memory optimization, kernel fusion, and parallel algorithm design.
• Comfortable from kernel to system-level infra.

Responsibilities

• Design, implement, and optimise CUDA kernels for high-throughput workloads.
• Profile GPU performance across compute, memory, and interconnect bottlenecks.
• Build tooling to correlate GPU power draw with energy pricing and grid signals.
• Optimise multi-GPU/multi-node scaling using NCCL, MPI, or similar libraries.
• Collaborate with data center teams on power capping and DVFS strategies.
• Integrate custom kernels into ML training/inference pipelines.
• Benchmark against CPU/GPU baselines and drive improvements.
• Contribute to internal libraries, docs, and best practices for GPU performance.

Skills

CUDA programming

C++

Python tooling

GPU architecture knowledge

Performance profiling

Multi-GPU scaling

Memory optimization

Energy-aware compute

Tools

Nsight Systems

NCCL

MPI

RDMA/InfiniBand

Job description

The Opportunity

Demand for high-performance compute capacity across the markets we operate in significantly outpaces what we can currently build, meaning speed to power and reliability are critical to how we scale. This puts CUDA/GPU performance engineering at the center of how Fuse scales its compute infrastructure.

Responsibilities

• Design, implement, and optimise CUDA kernels for high-throughput, latency-sensitive workloads.
• Profile and tune GPU performance across compute, memory bandwidth, and interconnect (NVLink/PCIe) bottlenecks.
• Build tooling to correlate GPU cluster power draw and utilisation with real-time energy pricing and grid signals.
• Optimise multi-GPU and multi-node scaling using NCCL, MPI, or similar communication libraries.
• Work with data center infrastructure teams on power capping, dynamic voltage/frequency scaling, and workload scheduling strategies that reduce energy cost and carbon intensity.
• Collaborate with ML/systems engineers to integrate custom kernels into training/inference pipelines.
• Benchmark against CPU/GPU baselines and drive continuous performance improvements.
• Contribute to internal libraries, documentation, and best practices for GPU performance engineering.

• 4+ years of experience writing production CUDA code, or equivalent strong project/industry experience.

• Deep understanding of GPU architecture (SMs, warps, memory hierarchy, occupancy).
• Proficiency in C++ and CUDA; experience with Python for tooling/orchestration.
• Experience with performance profiling tools (Nsight Systems/Compute).
• Familiarity with multi-GPU/multi-node scaling (NCCL, MPI, RDMA/InfiniBand).
• Strong grasp of memory optimisation, kernel fusion, and parallel algorithm design.
• Comfortable working across the stack from low-level kernels to system-level infrastructure.

Nice to Have

• Experience with Triton, cuDNN, cuBLAS, or custom ML inference/training frameworks.
• Exposure to data center power/thermal management or demand-response systems.
• Background in HPC, quantitative finance, or large-scale distributed systems.
• Familiarity with Kubernetes/Slurm for GPU cluster orchestration.
• Interest or experience in energy markets, grid systems, or sustainability-focused compute.

• Competitive salary and an equity sign-on bonus.
• Biannual bonus scheme.
• Fully expensed tech to match your needs.
• Breakfast and dinner allowance for office based employees.