Ruijia Yang | Efficient LLM Systems

I am a PhD student in Data Science and Analytics at the Hong Kong University of Science and Technology (Guangzhou), advised by Prof. Zeyi Wen. I received a B.Eng. in Computer Science from Sun Yat-sen University in 2024.

My research focuses on efficient LLM training and heterogeneous systems. I build memory-efficient systems for full-parameter fine-tuning across CPUs, GPUs, and storage. My publicly released work includes SlideFormer for single-GPU execution and SlideDP for multi-GPU scaling.

Email · GitHub · CV

Selected publications

Preprint

SlideDP: Scaling Host-Resident LLM Fine-Tuning Across Multiple GPUs

Ruijia Yang, Shiyuan Lin, Yulong Ao, Zhiyu Li, Yingli Zhao, Xianduo Li, Yonghua Lin, and Zeyi Wen

arXiv preprint arXiv:2609.34162, Sep 2026

arXiv Bib PDF Code

@article{slidedp,
  title = {SlideDP: Scaling Host-Resident LLM Fine-Tuning Across Multiple GPUs},
  author = {Yang, Ruijia and Lin, Shiyuan and Ao, Yulong and Li, Zhiyu and Zhao, Yingli and Li, Xianduo and Lin, Yonghua and Wen, Zeyi},
  year = {2026},
  month = sep,
  journal = {arXiv preprint arXiv:2609.34162}
}

DAC

An Efficient Heterogeneous Co-Design for Fine-Tuning on a Single GPU

Ruijia Yang and Zeyi Wen

In Proceedings of the 63rd ACM/IEEE Design Automation Conference (DAC), 2026

DOI arXiv Bib HTML PDF Code

@inproceedings{slideformer,
  title = {An Efficient Heterogeneous Co-Design for Fine-Tuning on a Single GPU},
  author = {Yang, Ruijia and Wen, Zeyi},
  year = {2026},
  doi = {10.1145/3770743.3804125},
  pages = {1--7},
  publisher = {Association for Computing Machinery},
  booktitle = {Proceedings of the 63rd ACM/IEEE Design Automation Conference (DAC)}
}

What's new

Sep 28, 2026	SlideDP is now on arXiv: scaling host-resident LLM fine-tuning across multiple GPUs.
Mar 17, 2026	Our DAC 2026 paper on SlideFormer is available on arXiv.