About Me
Hi, I am Shenggui Li, you can call me Frank Lee as well. I am a second-year PhD candidate at the Nanyang Technological University in Singapore, advised by Prof. Tianwei Zhang and Prof. Ivor Tsang. I am also affliated with Singapore A*STAR Center for Frontier AI Research (CFAR). Prior to that, I obtained my 🤓 B.Eng in Computer Science in Nanyang Technological University in 2021 advised by Prof. Chen Change Loy. I also worked as a founding member of a MLSys startup from 2022 to 2024.
My interest lies at the intersection of 🤖 machine learning and ⚡️ high performance computing. I first proposed the concept of sequence parallelism in 2021, which has attracted follow-ups and is widely adopted for large-scale model training. I am currently leading the SpecForge project and the SpecBundle intiative. I am also the author of Ring Self-Attention, Colossal-AI and Open-Sora. During my undergraduate study, I had the privilege to spend a wonderful year on the NTU HPC Team and we won the overall 🥈 2nd place in the student cluster competition of ISC21.
You can find my 📄 CV here.
I am currently working on builiding efficient systems for generative AI and write technical blogs sometimes. Feel free to contact me for collaboration.
During my spare time, I am learning and contributing to open-source projects such as
- 🚀 SGLang-Omni: Project Lead
- 🚀 SpecForge: Project Lead
- 🚀 SGLang: Core Dev
I was previously involved in some interesting projects.
- 🎥 Open-Sora: System Lead (until July 2024)
- ⚡️ SwiftInfer: Project Lead (until Jan 2024)
- ⚡️ ColossalAI: Project Lead, Top 1 contributor (until Jan 2024)
News
- [2026-05] 🎉 Three papers got accepted to ICML 2026, thanks to all my collaborators!
- [2026-04] 💻 Started internship at ByteDance Seed, working on LLM model architecture and efficiency.
- [2026-03] 🎉 One second-authored paper got accepted to ICS 2026.
- [2026-02] 🎉 One paper got accepted to MLSys 2026, congratulations to Qiaoling!
- [2026-02] 🎉 One first-authored paper got accepted to ICLR 2026.
- [2026-01] 🎉 One paper got accepted to ASPLOS 2026, congratulations to Runyu!
- [2025-07] 🎉 Released SpecForge v0.2
and 🤗 SpecBundle phase 1! Check out our blog !
- [2025-07] 🎉 Released SpecForge v0.1! Check out our blog !
- [2024-08] 🏫 Started my PhD journey at NTU CCSD, under the supervision of Prof Tianwei Zhang.
- [2024-06] 🎥 Released Open-Sora V1.2
with features such as 3D-VAE, rectified flow, and score conditio and better video generation quality!
Selected Publications
- Sequence Parallelism: Long Sequence Training from System Perspective
Shenggui Li, Fuzhao Xue, Chaitanya Baranwal, Yongbin Li, Yang You,
ACL 2023
- Colossal-AI: A Unified Deep Learning System For Large-Scale Parallel Training
Shenggui Li, Hongxin Liu, Zhengda Bian, Jiarui Fang, Haichen Huang, Yuliang Liu, Boxiang Wang, Yang You,
ICPP 2023
- SpecForge: A Flexible and Efficient Open-Source Training Framework for Speculative Decoding
Shenggui Li, Chao Wang, Yikai Zhu, Yubo Wang, Fan Yin, Shuai Shi, Yefei Chen, Xiaomin Dong, Qiaoling Chen, Jin Pan, Ji Li, Yineng Zhang, Lei Yu, Yonggang Wen, Ivor Tsang, Tianwei Zhang,
ICML 2026
- DSA: Efficient Inference For Video Generation Models via Distributed Sparse Attention
Shenggui Li, Runyu Lu, Qiaoling Chen, Haiyan Yin, Yueming Lyu, Yonggang Wen, Ivor Tsang, Tianwei Zhang,
ICLR 2026
- TetriServe: Efficiently Serving Mixed DiT Workloads
Runyu Lu, Shiqi He, Wenxuan Tan, Shenggui Li, Ruofan Wu, Jeff J Ma, Ang Chen, Mosharaf Chowdhury,
ASPLOS 2026
- DSB: Dynamic Sliding Block Scheduling for Diffusion LLMs
Lizhuo Luo*, Shenggui Li*, Yonggang Wen, Tianwei Zhang,
ICML 2026
- Online evolutionary batch size orchestration for scheduling deep learning workloads in GPU clusters
Zhengda Bian, Shenggui Li, Wei Wang, Yang You,
SC 2021
Honours and Awards
- [2024] A*STAR Computing & Information Science Scholarship
- [2021] Overall 2nd Place, ISC21 Student Cluster Competition
- [2020] Merit Award, APAC HPC-AI Competition
- [2018] NTU President Research Scholar
- [2017-2021] NTU Science & Technology Undergraduate Scholarship
- [2012-2016] Singapore SM1 Scholarship
