Remote CUDA GPU Software Engineer – HPC & AI

Bright Vision Technologies; United States; Remote; USD 100,000 - 175,000; Full time

Job summary

Bright Vision Technologies is seeking a GPU Software Engineer with deep CUDA expertise to design and optimize compute-intensive workloads on modern accelerator hardware. The role emphasizes delivering high-performance kernels for AI training and inference, with close collaboration across product, design, and engineering teams.

The ideal candidate has 10+ years of GPU programming experience, strong C++ skills, and a proven track record of production-grade performance improvements on GPU systems.

Qualifications

• Ten+ years in GPU programming and performance engineering.
• Deep expertise in CUDA C/C++ and GPU architectures.
• Experience profiling GPU workloads in production.

Responsibilities

• Design and implement high-performance CUDA kernels for compute-intensive workloads.
• Profile and optimize GPU code using Nsight tools and CUDA profilers.
• Tune memory patterns, occupancy, and shared memory for peak perf.
• Develop optimized libraries for linear algebra and ML primitives.
• Optimize multi-GPU training using NCCL and high-performance networking.
• Integrate custom kernels into ML frameworks.
• Collaborate with ML engineers to identify performance bottlenecks.
• Create benchmarks and regression tests to safeguard performance.
• Stay updated with CUDA evolution and accelerator technologies.

Skills

CUDA programming

GPU architecture

Performance profiling

C++ programming

Parallel computing

ML primitives

Education

Bachelor’s or Master’s in Computer Science/Engineering

Tools

Nsight Systems

Nsight Compute

NCCL

MPI

CUDA toolkit

Job description

Bright Vision Technologies is seeking a GPU Software Engineer with deep CUDA expertise to design and optimize compute-intensive workloads on modern accelerator hardware. The role emphasizes delivering high-performance kernels for AI training and inference, with close collaboration across product, design, and engineering teams.

The ideal candidate has 10+ years of GPU programming experience, strong C++ skills, and a proven track record of production-grade performance improvements on GPU systems.