Designing Rebellions' proprietary compiler stack to accelerate deep learning models on RBLN NPU products
Architecting and developing production-quality frontend/backend compilers with a focus on generalization, encompassing a wide range of functional coverage and optimization
Evaluating and verifying core-level and system-level functional features, working closely with hardware and system software architects/engineers
Key Qualifications
Master's or higher degree in Computer Science, Electrical Engineering, or a related field
Knowledge of compiler architecture, including various transformation passes, high/mid/low-level optimization/scheduling techniques, scratchpad/buffer memory allocation, backend code generation, etc.
Excellent troubleshooting, problem-solving, and debugging skills
Proficiency in programming languages: Python, C++
Ideal Qualifications
Proven track record of building high-quality, production-level software
Experience in deep learning inference on ASIC, including NPUs, GPUs, mobile APs, etc.