GPU Compiler Engineer
A
Apus Network
4 - 5K USD
Full-time
Remote
PythonC/C++GitGPUCUDALinux
- Design, develop, and maintain the clang/LLVM compiler toolchain for GPUs.
- Optimize GPU programs to improve performance.
- Communicate with hardware and software teams to resolve engineering issues.
- Develop and optimize GPU compilers (both front-end and back-end).
- Master's degree or higher in Computer Science, Electronic Engineering, Communications, or related fields, preferably from QS200 universities.
- 3+ years of experience in GPU compiler development (5+ years for senior engineers).
- Familiar with basic principles and optimization techniques of compilers.
- Proficient in C, C++, or Python.
- Proficient with software development tools such as Git, Linux, etc.
- Strong problem-solving and communication skills.
- Additional experience is a plus:
- Familiarity with CUDA PTX assembly instructions or AMDGPU assembly instructions.
- Familiarity with CUDA architecture and Tensor Core architecture.
- Familiarity with GPU programming models such as CUDA, HIP, OpenCL, OpenMP, etc.
- Familiarity with GPU thread synchronization, memory consistency, and related technologies.
- Experience optimizing GPU kernel programs.