Senior System Software Engineer, Cuda - Tegra

NVIDIA NVIDIA · Semiconductors · Bangalore, India

NVIDIA is seeking an experienced software engineer to contribute to the CUDA driver and runtime, a core component of NVIDIA's CUDA parallel computing platform. The role involves building and maintaining features, contributing to the full SDLC, writing production-level code, and collaborating with hardware teams on next-generation GPUs. Requires a BS/MS in CS or equivalent experience, strong C/C++ skills, OS fundamentals, and debugging abilities.

What you'd actually do

  1. Build and maintain features in CUDA for multiple NVIDIA platforms
  2. Contribute towards all aspects of SDLC, and collaborate across multiple teams
  3. Write effective, maintainable, and well-tested production level code
  4. Participate in the development of proprietary and open computing standards
  5. Contribute to development of next-generation GPUs in collaboration with HW teams at NVIDIA

Skills

Required

  • BS/MS in Computer Science or related field or equivalent 5+ years of demonstrated experience
  • System SW development life cycle experience
  • Strong knowledge of C/C++
  • good understanding of computer architecture and OS fundamentals (Linux, QNX or similar modern OS)
  • Excellent debugging and problem-solving skills
  • Self motivated with a passion to work on challenging problems in a dynamic environment
  • Great teammate with good communication and presentation skills
  • Ability to lead from the front, and work independently with minimal direction

Nice to have

  • Previous experience with CUDA
  • System Programming
  • SDLC
  • Expertise in Parallel Programming
  • Familiarity with AI tools

What the JD emphasized

  • 5+ years of demonstrated experience
  • System SW development life cycle experience
  • Strong knowledge of C/C++
  • good understanding of computer architecture and OS fundamentals (Linux, QNX or similar modern OS)
  • Excellent debugging and problem-solving skills
  • Ability to lead from the front, and work independently with minimal direction