Liangsheng Yin 「尹良升」

Hello, my name is Liangsheng Yin, and I am an undergraduate student at Shanghai Jiao Tong University, enrolled in ACM Honor Class. I am also a research assistant at Sky Computing in UC Berkeley. I am majoring in Computer Science, with a deep interest in the fields of MLSys, Distributed System and Networking.

Currently, I am fortunate to work with Lianmin Zheng and Ying Sheng from LMSYS.ORG, working as a core developer of SGLang, and avised by Ion Stoica and Joseph E. Gonzalez in Sky Computing. We are committed to developing more efficient and powerful system for Artificial Intelligence. I am also an applicant for the 2025 Fall Ph.D. program in Computer Science.

Feel free to check out my CV and drop me an e-mail if you want to chat with me!

~ Email | CV | Google Scholar | Github | Twitter ~

Jul '25	Super fast SGLang v0.2(v.s. TensorRT, vLLM) is released! Check it out here.
Jul '04	Arrived at UC Berkeley and thrilled to start my journey with Sky Computing. Feel free to reach out!
Feb '05	The compressed FSM , a new feature for Faster JSON/regex decoding is available in SGLang.

	Research Assistant \| Sky Computing at UC Berkeley July '24 - Present Working with Ion Stoica and Joseph E. Gonzalez form Sky Computing and LMSYS Group.
	Research Intern \| Large Model Systems Organization September '23 - Present Working with Lianmin Zheng form UC Berkeley and Ying Sheng from Stanford University. We are commited to developing large models and systems that are open, accessible, and scalable.
	Undergraduate Student \| Shanghai Jiao Tong University September '21 - Present Working under the supervision of Prof. Yong Yu and majoring in Computer Science, expected to graduate in 2025.

SGLang: Efficient Execution of Structured Language Model Programs
[preprint] [code]

Lianmin Zheng*, Liangsheng Yin, Zhiqiang Xie, Jeff Huang, Chuyue Sun, Cody Hao Yu, Shiyi Cao, Christos Kozyrakis, Ion Stoica, Joseph E. Gonzalez, Clark Barrett, Ying Sheng*

	Fast JSON Decoding for Local LLMs with Compressed Finite State Machine [blogpost] [code] Liangsheng Yin, Ying Sheng, Lianmin Zheng
	Achieving Faster Open-Source Llama3 Serving with SGLang Runtime (vs. TensorRT-LLM, vLLM) [blogpost] Liangsheng Yin, Yineng Zhang, Ying Sheng et al. in SGLang team

ChatGPT-Web

This web is a modification to Rishab Khincha's website.