Software Engineer (GPU)

TL;DR

Software Engineer (GPU): Developing and optimizing GPU-accelerated software for internal tooling and customer solutions with an accent on CUDA and HIP programming, kernel-level improvements, and multi-vendor GPU performance tuning. Focus on profiling, debugging, and validating GPU workloads across NVIDIA and AMD architectures to deliver production-grade, high-performance code.

Location: San Jose, California, United States

Salary: $153,000 - $167,000

Company

Top tier provider of advanced server, storage, and networking solutions for data center, cloud computing, enterprise IT, and IoT/embedded customers worldwide.

What you will do

• Implement and optimize GPU-accelerated code using CUDA, HIP, and vendor SDKs.
• Port workloads between NVIDIA and AMD GPU platforms with minimal regression.
• Develop benchmarking suites, diagnostics, and performance tooling.
• Collaborate with hardware engineering and solution architects to align software with system constraints.
• Profile, debug, and validate GPU workload performance using specialized tools.
• Maintain documentation and contribute to continuous integration pipelines for GPU builds and tests.

Requirements

• Location: Must be based in San Jose, California, United States
• 3-5 years experience with CUDA, HIP, GPU compute fundamentals, and parallel programming.
• Proficiency in C/C++ and Python with Linux development experience.
• Experience with GPU profiling tools such as Nsight, rocprof, and OmniPerf.
• BS in EE/CS/CIS or related field; Master’s preferred.
• English proficiency at least B2 level.

Nice to have

• Experience with ROCm, Triton, OpenCL, or domain-specific kernel DSLs.
• Knowledge of GPU virtualization, containerized execution, and multi-GPU communication.
• Background in system-level performance bottlenecks and mixed-precision strategies.