Senior Performance Compiler Engineer - Triton
Core
Design and implement compiler technology using MLIR to optimize high-level kernel descriptions for NVIDIA GPUs, enabling breakthroughs in large language models and AI applications.
Role type
Senior IC performance compiler engineer
Builds
Efficient, low-level GPU code for AI training and inference
Domain
AI computing, high-performance computing, compiler technology
Deliverable
production ML models
Required skills
C++ programming, software design, performance analysis, parallel programming (CUDA/OpenCL), computer architecture, assembly-level programming
Preferred skills
BLAS tuning, deep learning library kernels, numerics and linear algebra, MLIR/TVM experience, open-source contributions
Technologies
MLIR, Triton, CUDA, PTX, OpenCL, OpenMP
Responsibilities
Investigate NVIDIA GPU hardware architecture and programming models; understand advanced AI algorithms and numerics to identify optimization opportunities; design and implement compiler technology to generate efficient GPU code; engage in iterative optimization processes; collaborate with hardware architects and CUDA compiler teams
Seniority
Senior, hands-on IC
