2027AIDU-AI异构计算研发工程师(J99964)
Core
Developing AI heterogeneous computing acceleration engines and platforms for inference and training, optimizing algorithms for cloud computing, search, and multimodal AI tasks.
Role type
Senior IC heterogeneous computing software engineer
Builds
AI inference engines, communication libraries, and heterogeneous high-performance computing platforms
Domain
AI infrastructure, high-performance computing, heterogeneous architectures
Deliverable
production ML models
Required skills
C++, Python, GPU/ARM/MIPS/DSP heterogeneous platforms, computer architecture, assembly-level development, MPI, RDMA, GPU direct, distributed computing strategies
Preferred skills
PaddlePaddle/TensorFlow source optimization, CUDA/OpenCL with SASS/PTX optimization, Neon/ARM-GPU development, hardware performance analysis toolchains, LLVM, Linux kernel development
Technologies
CUDA, OpenCL, MPI, NCCL, RDMA, GPU Direct, LLVM, Neon, ARM-GPU, CodeXL, NVVP, GPA
Responsibilities
Develop and optimize large-scale AI inference engine frameworks and underlying operators; develop and optimize AI computing communication libraries and algorithms; develop compilation systems and optimize algorithms for CPU/GPU/FPGA/ASIC architectures; design and develop heterogeneous high-performance computing platforms and libraries
Seniority
Senior, hands-on IC
