About the job
The Computing Data Application Acceleration Lab at Huawei Canada in Markham works on global data analytics platforms. This team combines software and hardware expertise to improve data efficiency in storage and runtime. Projects include developing advanced GPU architectures for gaming, cloud rendering, VR/AR, and Metaverse applications, always with future needs in mind. The group aims to advance algorithm performance and training efficiency across industries to support long-term competitiveness.
Role overview
This internship centers on research and development for AI foundation model training. The focus is on supporting and optimizing large language models (LLMs), code models, and multimodal models. The work involves model architecture improvements, running experiments, applying post-training optimization techniques, and exploring continual learning. Hardware-aware strategies to boost model efficiency are part of the scope.
Key responsibilities
- Assist in developing and refining foundation models, including architecture improvements and experimental work.
- Apply post-training optimization and investigate continual learning methods.
- Help build and improve distributed training and inference systems, including parallelization strategies such as model, tensor, and data parallelism.
- Support operator-level and computational graph optimizations, and participate in performance benchmarking and analysis.
- Work closely with hardware architects and algorithm engineers on research, experiments, and prototyping to enhance system and model performance.
Compensation
Total target annual compensation (based on 2,080 hours per year) ranges from $58,000 to $104,000, depending on education, experience, and demonstrated expertise.
