GPU HPC Engineer — CUDA/C++ Performance Expert

Hire Feed; South Africa; À distance; ZAR 600 000 - 900 000; Plein temps

Résumé du poste

Hire Feed is seeking a GPU Programming Software Engineer to join a client project on a full-time basis. You will develop high-performance computing solutions leveraging GPU architectures and optimize GPU-accelerated workloads.

You will design CUDA/OpenCL kernels, optimize parallel code, and collaborate with cross-functional teams to deliver scalable GPU solutions for large-scale data processing. Strong C/C++ skills and a deep understanding of GPU libraries are essential.

Qualifications

• Proficiency in CUDA programming and GPU optimization.
• Experience with OpenCL, HIP or other GPU frameworks.
• Strong C/C++ programming skills focused on performance.
• Familiarity with parallel algorithms and data structures.
• Knowledge of cuBLAS, cuFFT, or Thrust libraries.

Responsabilités

• Design GPU-optimized software for HPC applications.
• Develop CUDA/OpenCL kernels and libraries.
• Profile and optimize code to improve GPU utilization and latency.
• Collaborate with teams to integrate GPU acceleration.
• Write unit tests and benchmarks to validate performance.

Connaissances

CUDA programming

C/C++ programming

parallel algorithms

GPU architecture optimization

Outils

OpenCL

HIP

cuBLAS

cuFFT

Thrust

Description du poste

Hire Feed is seeking a GPU Programming Software Engineer to join a client project on a full-time basis. You will develop high-performance computing solutions leveraging GPU architectures and optimize GPU-accelerated workloads.

You will design CUDA/OpenCL kernels, optimize parallel code, and collaborate with cross-functional teams to deliver scalable GPU solutions for large-scale data processing. Strong C/C++ skills and a deep understanding of GPU libraries are essential.