Lianmin Zheng
@lm_zheng
Followers
12K
Following
2K
Media
3
Statuses
375
Member of technical staff @xAI | Prev: Ph.D. @UCBerkeley, Co-founder @lmsysorg
Bay Area, California
Joined January 2018
RT @rajko_rad: We @a16z just launched the third batch of Open Source AI Grants (cc @mbornstein) 🎉. This round includes projects focused on….
0
58
0
RT @NVIDIAAIDev: .@lmsysorg (SGLang) now achieves 7,583 tokens per second per GPU running @deepseek_ai R1 on the GB200 NVL72, a 2.7x leap o….
0
36
0
RT @simon_ycl: Introducing RL2: Ray-Less Reinforcement Learning for LLMs 🚀 . Want to run RL experiments but tired of complicated abstractio….
0
36
0
RT @jxwuyi: We release fully async RL system AReaL-boba² for LLM & SOTA code RL w. Qwen3-14B! @Alibaba_Qwen #opensource .🚀system&algorithm….
0
40
0
RT @bfspector: (1/5) We’ve never enjoyed watching people chop Llamas into tiny pieces. So, we’re excited to be releasing our Low-Latency-L….
0
144
0
Blackwell-specific optimizations are cooking! 🚀.
We are incredibly grateful to @nvidia @NVIDIAAIDev for the generous donation of a DGX B200 to SGLang, in addition to a previous donation of DGX H100! SGLang cannot move such fast without NVIDIA’s support.❤️. We also thank our data center partner InnoMatrix for hosting the
4
5
74
RT @novita_labs: Thrilled to partner with @lmsysorg! 🚀. Our high-performance GPU cloud is powering their game-changing inference engine, re….
0
7
0
RT @ye_combinator: We’re thrilled that FlashInfer won a Best Paper Award at MLSys 2025! 🎉.This wouldn’t have been possible without the comm….
0
37
0
RT @yi_xin_dong: We are hosting a happy hour with @lmsysorg at #mlsys2025! Join us for engaging talks on SGLang, the structured generation….
0
12
0
RT @zhyncs42: I’ll be joining my @basetenco colleague @philip_kiely at the AI Engineer World’s Fair @aiDotEngineer in San Francisco, June 3….
0
6
0
RT @skypilot_org: E2E guide to self-host multi-node Llama 4 with one command, using SkyPilot, SGLang (@lmsysorg), and `llm` (@simonw). ▸ S….
0
6
0
RT @lmsysorg: 🚀 Breaking: SGLang provides the first open-source implementation to serve @deepseek_ai V3/R1 models with large-scale expert p….
0
81
0