We are building the world’s most advanced, AI edge inference engine. We have core processor technology that is modular, low power and high performance among peer companies. We are now scaling our engineering team to build products to address various markets’ needs (Automotive, IOT etc.)
About the role:
AlphaICs is looking for an accomplished Compiler Engineer to design, implement and test new features for our compiler. In this role, closely with hardware designers and AI experts to identify and deploy support for key hardware capabilities.
Day-to-day responsibilities include:
- Design, develop, enhance, optimize AlphaICs a deep learning compiler
- Integrate compiler, code generation tools, run time tools to the main SDK.
- Define and develop compiler optimization for AlphaICs compiler.
- Collaborate closely with teams developing other related components to ensure compatibility, robustness, and high-quality code generation.
- Provide design documentation to software and platform teams.
- Participate in design reviews for software systems.
You should apply for this role if you have the following qualifications:
- BS, MS or Ph.D. in Computer Science, Computer Engineering, or related fields, or equivalent experience.
- A minimum of 4+ years of experience in compiler back end/front end, programming language designs, Compilers/Linkers.
- Hands-on experience in design and development of graph based compilers and experience in LLVM, MLIR, GLOW etc is a plus.
- Knowledge of CPU/GPU pipeline structure & memory organization.
- Experience in Deep Learning software stacks and architectures is a big plus.
- Knowledge and experience in low level programming and CPU or GPU performance optimizations, including parallel/distributed algorithms/DSP is desirable.
- Familiarity with hardware architectures beyond CPU and GPU (e.g., TPU, VPU, FPGA, etc.).
- Excellent programming/debugging skills at the Assembly level.
- Experience with parallel programming domain OpenMP is a plus.
- Experience in parallel computing domain like (multi-core, GPGPU, SIMD, MIMD etc) is a plus.