Principal Scientist - Software/Hardware Co-design

About the role

Huawei Canada’s Computing Data Application Acceleration Lab is seeking a Principal Scientist to lead software‑hardware co‑design initiatives. The role focuses on building universal AI performance models, tracking emerging hardware trends, and driving algorithm‑hardware breakthroughs for next‑generation AI chips.

Key responsibilities

• Develop accurate AI performance models across mainstream acceleration technologies to support theoretical analysis.
• Monitor and analyze emerging hardware designs, providing insight and direction for cutting‑edge technologies.
• Collaborate with AI research teams to pinpoint performance bottlenecks and define co‑design features for future chips.
• Perform quantitative performance modelling of representative AI workloads on varied hardware specifications.
• Lead the team in achieving optimal trade‑offs between model quality and compute efficiency.
• Stay abreast of algorithm‑hardware co‑design advances and synthesize industry trends.

Required profile

• Master’s or Ph.D. in Computer Science, Electronic Engineering, or a related field.
• 5+ years of experience in low‑level computing algorithm development, AI accelerator or large‑scale parallel/high‑performance computing system design.
• Deep knowledge of large language and multimodal models, AI software stacks (operators, compilers, acceleration libraries, frameworks), and modern training/inference techniques.
• Familiarity with AI chip micro‑architecture is a strong asset.

Required skills

• AI performance modelling and benchmarking.
• Algorithm‑hardware co‑design expertise.
• Low‑level computing algorithm development.
• Parallel and high‑performance computing.
• Understanding of AI software stack components (operators, compilers, acceleration libraries, frameworks).
• Experience with large language model optimization techniques such as hybrid parallelism, low‑precision formats, sparsity, and pipeline/data parallel splitting.