Yang Sui

Yang Sui is a Postdoctoral Research Associate in the Department of Computer Science at Rice University, collaborating with Professors Xia (Ben) Hu and Hanjie Chen. His research focuses on Efficient AI and Trustworthy AI, including deep neural networks, large language models (LLMs), diffusion models, and algorithm-hardware co-design. He holds a PhD from Rutgers University (2024), an MS from Jilin University (2019), and a BS from Jilin University (2016).

Education:

• PhD, Computer Science, Rutgers University, 2024
• MS, Computer Science, Jilin University, 2019
• BS, Computer Science, Jilin University, 2016

Research Interests:
Efficient AI: Model Compression (pruning, quantization, low-rank decomposition), Generative AI (diffusion models, LLMs), and algorithm-hardware co-design. Trustworthy AI: Adversarial robustness (backdoor attacks, vulnerability detection). He has interned at Snap Research (2024), Tencent America (2022), and Baidu (2018), contributing to projects like BitsFusion quantization and Paddle-Lite framework.

Awards:

• Paul Panayotatos Scholarship (2024)
• Best Paper Runner-Up Award (DCAA Workshop at AAAI 2023)
• First Place in ESWEEK Classification Track (2023)
• SGS Travel Award (2023)

Advising & Grants:
Advises students on topics like LLM quantization and multimodal models. Collaborates with industry and academia on grants related to efficient AI and hardware co-design. Led projects like Rice’s “Efficient Deep Learning Reading Group” (2023).

Labs & Teams:
Contributes to Snap’s Creative Vision team, Rutgers’ research groups, and co-design initiatives with industry partners like Baidu and Tencent.