Senior Software Engineer, CUDA UMD - Graphs and GPU Sharing

Role responsibilities

The role involves architecting and implementing new features for the CUDA Driver to optimize GPU performance for diverse computational workloads. Key tasks include extending CUDA Graphs and MPS while collaborating across teams to define future API improvements.

Requirements

Candidates need a BS or MS in Computer Science or Electrical Engineering with 8-10 years of development experience in C/C++. Proficiency in operating system interfaces, multithreaded programming, and large codebase management is required.

Key skills

C, C++, CUDA, Device Drivers, Multithreading, Operating Systems, Virtual Memory, Parallel Computing, System Architecture, Kernel Mode Development, Linux Systems Software, CUDA Graphs, Multi-Process Service (MPS), Memory Coherence, Memory-Mapped IO, Interconnects

Keywords

CUDA, GPU, C, C++, Device Drivers, Parallel Computing, Deep Learning, Scientific Computation, Data Science, CUDA Graphs, MPS, Linux, Kernel Mode, Virtual Memory, Multithreading, System Architecture, Memory Hierarchy, Interrupts, Memory-Mapped IO, Memory Coherence, Consistency Models, Software Engineering, Operating Systems