sayzardFeb 26

DeepReinforce (@deep_reinforce)

IterX가 MLSys 2026 NVIDIA 트랙의 Fused MoE 설정에서 FlashInfer AI 커널 생성 대회 벤치마크로 H100에서 15.62×, B200에서 14.84× 속도 향상을 달성해 GPT-5.2 Pro와 Claude 4.6 Opus를 크게 앞섰습니다. 참가자 전원에 무제한 크레딧을 제공한다고 공지했습니다.

https://x.com/deep_reinforce/status/2026849608828334337

#iterx #mlsys #fusedmoe #nvidia #flashinfer

DeepReinforce (@deep_reinforce) on X

🥳IterX for MLSys 2026 NVIDIA Track Fused MoE 🧑‍🍳IterX achieves a 15.62× speedup on H100 and 14.84× on B200, significantly surpassing GPT-5.2 Pro and Claude 4.6 Opus on the Fused MoE setting of FlashInfer AI Kernel generation contest. 🎁We offer unlimited credits for all

X (formerly Twitter)
Shawn HymelFeb 28, 2023
We got accepted into #MLSys!
---
RT @EdgeImpulse
Our paper, "Edge Impulse: An MLOps Platform for Tiny Machine Learning," has been accepted at MLSys 2023! https://bit.ly/3IXCYqk
https://twitter.com/EdgeImpulse/status/1630573517858111492
Edge Impulse Partners with Harvard SEAS to Release Academic Paper

Our paper details the steps taken during a typical edge ML project, from data collection and analysis to model training and deployment.

marcoJan 15, 2023

An overview of Flash Attention at the Stanford #MLSys seminar:

https://youtu.be/gMOAud7hZg4

#ML #GPU #Transformers

FlashAttention - Tri Dao | Stanford MLSys #67

YouTube
Markus WeimerNov 21, 2022

#Introduction

I am a computer scientist interested in making #MachineLearning (#ML) more useful to more people.

So far, I worked on research in #ML and #Systems (#MLSys) for those to become "more useful". More recently, I focus on "more people". I worked on ML.NET to bring ML to the #dotnet platform. Hope to share more recent work soon :)

I support OSS. ML.NET is OSS. I am a Member of The Apache Software Foundations #ASF where I have contributed to REEF, TVM, mxnet, and Nemo in varying roles.