LMSYS Org announces talks on AI agents tuning SGLang serving and GPU kernels
Overview
LMSYS Org says AMD's Emad Barsoum will show that agents can profile the stack, run automated experiments, and tune SGLang serving configs and GPU kernels.
The post announcing the talk gives no figures.
In a separate LMSYS Org post, Chengze Fan and Oguz Ulgen from Meta are listed to present "Agentic Development of Megakernels on AMD," on using coding agents to build GPU kernels over days. The talk covers profiling and numerical parity checks for faster serving.
Written by AI from the articles below · updated Nov 9, 4:12 PM ET
Check the sources:
Developments
2 developments
- Nov 8, 8:58 PM ET · 1 articleMeta engineers present agentic development of megakernels on AMDLMSYS Org: Meta engineers present agent-built GPU megakernels on AMD hardware
- Nov 8, 8:58 PM ET · 1 articleEmad Barsoum of AMD shows agents tuning SGLang serving and GPU kernelsLMSYS Org: Emad Barsoum of AMD shows agents tuning SGLang serving and GPU kernels
Article timeline
The articles in this story. Times are ET.
LMSYS Org@lmsysorgOfficialMeta engineers present agent-built GPU megakernels on AMD hardwareAIChengze Fan and Oguz Ulgen from Meta present "Agentic Development of Megakernels on AMD," on using coding agents to build GPU kernels over days. The talk covers profiling and numerical parity checks for faster serving.

LMSYS Org@lmsysorgOfficialEmad Barsoum of AMD shows agents tuning SGLang serving and GPU kernelsAIEmad Barsoum of AMD says agents can profile the stack, run automated experiments, and tune SGLang serving configs and GPU kernels. The post is a short announcement of the talk, with no figures given.

Heat Trend
- Current comparable heat
- 5
- Peak comparable heat
- 6Nov 9, 5:00 PM ET
- Change in past 24 hours
- –
The trend compares only continuously observed participants, so its scope may be smaller than the current heat total. Hover or tap for hourly heat; use arrow keys with a keyboard.