Negotiable
Outside
Remote
USA
Summary: The Hub-Compiler Developer will play a crucial role in designing and developing software tools for a digital neuromorphic accelerator chip. This position involves collaborating with other software developers under the guidance of the client's team lead. The ideal candidate will possess expert-level programming skills and experience in optimizing code for custom hardware. Responsibilities include developing software across various levels of the stack, from compilers to automation tools.
Key Responsibilities:
- Develop optimized code for custom hardware to implement neural network layers
- Develop at all levels of software stack (compiler, runtime, package and test automation)
Key Skills:
- Expert-level knowledge of C++, Python and NorthPole microcode
- Shell scripts (bash)
- Build systems (Makefile, CMake) and continuous integration tools (Docker, Jenkins)
- Source control (Git) and collaboration tools (Confluence, JIRA)
Salary (Rate): undetermined
City: undetermined
Country: USA
Working Arrangements: remote
IR35 Status: outside IR35
Seniority Level: undetermined
Industry: IT
Statement of Work:
The engineer will assist in the design and development of the software toolchain for a digital neuromorphic accelerator chip. The selected candidate will work directly with other software developers under the supervision of client's team lead.
Task Description:
- Develop optimized code for custom hardware to implement neural network layers
- Develop at all levels of software stack (compiler, runtime, package and test automation)
Required skills/Level of Experience :
- Expert-level knowledge of C++, Python and NorthPole microcode
- Shell scripts (bash)
- Build systems (Makefile, CMake) and continuous integration tools (Docker, Jenkins)
- Source control (Git) and collaboration tools (Confluence, JIRA)
Nice to have skills:
- More than 4 years experience as a software developer
- Parallel programming (CUDA, OpenMP)
- Assembly / microcontroller / DSP programming
- Compilers (gcc, clang, LLVM)
- Neural network compilers (ONNX, TVM, TensorRT)
- Familiarity with neural network operations like convolution, pooling, recurrent networks
- Data visualization (OpenGL, Dash)
- Linux kernel/device drivers (Ubuntu, CentOS)
- Compilers (gcc, clang, LLVM)
- Parser generators (ANTLR)
- Parallel programming (CUDA, OpenMP)
- Neural network training frameworks (PyTorch, TensorFlow)
- Neural network compilers (ONNX, TVM, TensorRT)