Overview
The contractor will join a specialized team focused on developing high-performance GPU-accelerated software for scientific and AI applications. Working remotely, the CUDA Developer will collaborate with C++ engineers and machine learning researchers, contributing to architecture decisions and optimizing CUDA kernels for computational efficiency.
Responsibilities
- Design and optimise CUDA kernels for high-performance workloads.
- Translate advanced algorithms into production-ready GPU-accelerated code.
- Profile performance and reduce bottlenecks using Nsight, CUPTI, and custom tooling.
- Collaborate with C++ engineers and ML researchers to deliver scalable AI computation pipelines.
- Contribute to architecture decisions on parallelisation, data transfer, and memory efficiency.
Requirements
- Deep experience with CUDA C/C++ and modern C++ (C++17/20).
- Strong understanding of GPU architecture, memory management, and parallelism.
- Familiarity with OpenMP, MPI, or other HPC frameworks.
- Exposure to AI/ML workloads or scientific computing is a plus.
- Ability to work collaboratively in fast-moving, high-impact environments.