高性能计算研发工程师（J106580）
Core
Develop AI compilers and toolchains to enable efficient execution of deep learning models on AI accelerators, focusing on operator optimization, code generation, and performance modeling.
Role type
Senior IC AI compiler and high-performance computing engineer
Builds
AI compiler toolchains, optimized operators (e.g., matrix multiply, Attention), and performance simulation tools
Domain
AI infrastructure, compilers, and high-performance computing
Deliverable
production ML models
Required skills
C/C++, Python, compiler principles, computer architecture, parallel computing, deep learning operator analysis, LLM-assisted code generation
Preferred skills
LLVM/MLIR/TVM experience, GPU/NPU core operator development, performance modeling and microarchitecture simulation
Technologies
C/C++, Python, LLVM, MLIR, TVM, GPU, NPU
Responsibilities
Develop and optimize high-performance operators (matrix multiply, Attention); research Persistent Kernel and device-side task scheduling; build AI accelerator performance models and simulation tools; explore microarchitecture design spaces for hardware-software co-optimization
Seniority
Senior, hands-on IC