Senior Performance Software Engineer, Deep Learning Libraries
Core
Developing optimized code to accelerate linear algebra and deep learning operations on NVIDIA GPUs for cuDNN, cuBLAS, and TensorRT libraries.
Role type
Senior IC performance software engineer (deep learning libraries)
Builds
High-performance code for NVIDIA's cuDNN, cuBLAS, and TensorRT libraries
Domain
Deep learning, GPU computing, high-performance computing
Deliverable
production ML models
Required skills
C++, parallel programming, computer architecture, assembly programming, performance analysis, debugging, test design
Preferred skills
CUDA GPU programming, BLAS tuning, numerical methods, linear algebra, LLVM, TVM, TensorFlow MLIR
Responsibilities
Writing highly tuned compute kernels for core deep learning operations (matrix multiplies, convolutions, normalizations); Collaborating with CUDA compiler, deep learning performance, and hardware teams; Identifying bottlenecks and optimizing resource utilization to improve throughput.
