Senior CUDA Engineer for High-Performance Inference

Senior CUDA Engineer for High-Performance Inference

Full-Time 60000 - 80000 £ / year (est.) No working from home possible
Fuse Energy

At a Glance

  • Tasks: Design and optimise GPU code for high-performance inference workloads.
  • Company: Join Fuse Energy, a leader in innovative tech solutions.
  • Benefits: Competitive equity package and a vibrant London office.
  • Other info: Opportunity to take ownership of engineering quality and tests.
  • Why this job: Make a real impact on cutting-edge GPU technology and performance.
  • Qualifications: Experience with CUDA, GPU programming, and performance optimisation.

The predicted salary is between 60000 - 80000 £ per year.

Fuse Energy is seeking an experienced CUDA Engineer to design and optimise low‑level GPU code for high‑throughput inference workloads.

You will work on custom CUDA kernels, memory hierarchies, and mixed‑precision arithmetic to maximise throughput and minimise latency across GPU fleets.

The role focuses on profiling, kernel fusion, and performance‑driven optimisations with strong ownership of engineering quality and tests.

London‑based office, competitive equity package.

#J-18808-Ljbffr

Senior CUDA Engineer for High-Performance Inference employer: Fuse Energy

Fuse Energy is an exceptional employer, offering a dynamic work environment where innovation meets opportunity. As an AI Compute Data Centre Engineer, you will be at the forefront of cutting-edge technology in the UK, with access to competitive salaries, biannual bonuses, and fully expensed tech. Our culture fosters professional growth and collaboration, ensuring that every team member can thrive and contribute to impactful projects.

Fuse Energy

Contact Details:

Fuse Energy Recruitment Team

We think you need these skills to ace Senior CUDA Engineer for High-Performance Inference

CUDA Programming
GPU Optimisation
Low-Level GPU Code Design
Custom CUDA Kernels
Memory Hierarchies
Mixed-Precision Arithmetic
Throughput Maximisation