Principal Software Engineer, Kernels
Core
Developing and optimizing software kernels for next-generation AI hardware to productize the company's AI compute engine stack.
Role type
Principal Software Engineer, Kernels
Builds
Software kernels and compiler infrastructure for AI compute engines
Domain
AI hardware, computer architecture, systems software
Deliverable
production ML models
Required skills
Computer architecture, C/C++, Python, Linux, CUDA, FPGA/DSP/GPU/AI accelerator programming, ML operators (GEMM, Conv, BLAS, SIMD), embedded SIMD vector processors
Preferred skills
ML frameworks (TensorFlow, PyTorch), ML compilers (MLIR, LLVM, TVM, Glow), deep learning models (CV, NLP, recommendation), cloud provider or AI compute company experience
Technologies
CUDA, MLIR, LLVM, TVM, Glow, TensorFlow, PyTorch, Tensilica
Responsibilities
Develop, enhance, and maintain software kernels for AI hardware; map computational graphs to underlying hardware architectures; optimize hardware-software co-design trade-offs; build compiler infrastructure with compiler experts; collaborate with ML, systems, and hardware teams
Seniority
Principal, hands-on IC with leadership
