GPU Software Engineer (CUDA)

Role Overview

We are seeking a GPU Software Engineer (CUDA) to design and optimize compute-intensive workloads on modern accelerator hardware. The ideal candidate combines low-level systems mastery with strong software engineering practices, and has a track record of delivering measurable performance improvements on production GPU systems.

What You Will Do

Design and implement high-performance CUDA kernels, profile and optimize GPU code, tune memory access patterns, and develop highly optimized libraries for linear algebra and ML primitives.

Why It Might Be a Fit

The successful candidate brings strong engineering discipline, a clear communication style, and a track record of shipping meaningful work that holds up well in production.

Requirements

• Bachelor’s or Master’s degree in Computer Science, Computer Engineering, or a related field.
• Six or more years of experience in GPU programming and performance engineering.
• Deep expertise in CUDA C/C++ and GPU programming models.
• Strong understanding of modern GPU architectures, memory hierarchies, and execution models.
• Hands-on experience profiling and optimizing GPU workloads in production.
• Familiarity with NCCL, MPI, and high-performance interconnect technologies.
• Experience integrating custom kernels into ML frameworks.
• Strong C++ skills and familiarity with modern systems programming practices.
• Solid grounding in linear algebra and numerical methods.
• Strong communication and collaboration skills with research and engineering teams.

Benefits

• Competitive base salary commensurate with experience, plus benefits.
• 100% remote work option.
• Long-term, multi-year employment engagement.
• H1B transfers welcomed for qualified candidates.
• No new H1B sponsorship available.

Originally posted on Himalayas