Description
NVIDIA is hiring a Senior Performance Software Engineer for Deep Learning Libraries to develop highly optimized compute kernels for linear algebra and deep learning operations on NVIDIA GPUs. The role focuses on improving performance in cuDNN, cuBLAS, and TensorRT, supports regression testing and CI/CD, and collaborates with CUDA compiler, deep learning performance, and hardware teams. Candidates need a master’s or PhD in a relevant field, at least two years of relevant industry experience, strong C++ and parallel programming skills, computer architecture knowledge, and experience with assembly programming.
