Hongyi Wang (@HongyiWang10) / X

Hongyi Wang

858 posts

Hongyi Wang

@HongyiWang10

Assist. Prof. @RutgersCS; Head of Infra @genbioai; Ex @mldcmu @WisconsinCS

Piscataway, NJ

Joined August 2019

Pinned
Hongyi Wang
@HongyiWang10
Jun 2
1/8) Excited to share PR²: Predictive Routing Replay for MoE-Based LLM RL! 🎉 This is also the first paper from our research group at @RutgersCS @RutgersU. While MoE LLMs scale remarkably well, RL training exposes a hidden source of instability: routing. Paper:
Hongyi Wang
@HongyiWang10
Dec 16, 2024
I have three Ph.D. student openings in my research group at @RutgersCS starting in Fall 2025. If you are interested in working with me on efficient algorithms and systems for LLMs, foundation models, and AI4Science, please apply at: grad.rutgers.edu/academics/prog… The deadline is
Hongyi Wang
@HongyiWang10
May 24, 2024
[1/n] I'm thrilled to share that I will join the Rutgers CS Department @RutgersCS as a tenure-track Assistant Professor in the summer of 2025! I'm excited about and looking forward to this new chapter of my career journey!
Hongyi Wang
@HongyiWang10
Oct 22, 2023
1/ I am currently on the academic job market, applying for Assistant Professor positions in any field related to CS! My research focuses on ML & Systems, specifically on computation- and communication-efficient distributed ML, efficient computing in LLMs, and federated learning.
Hongyi Wang
@HongyiWang10
Oct 4, 2023
1/ Our paper is out! We developed the Fusion of Experts (FoE) that merges expert model and LLM outputs for better performance across image & text classification, text summarization, MMLU, and automatic text evaluation. paper: arxiv.org/abs/2310.01542 Thread below.
Hongyi Wang
@HongyiWang10
Dec 19, 2024
My team (i.e., the AI Infra Team) @genbioai is hiring! If you’re passionate about developing: 1. Large-scale AI systems, apply here: jobs.lever.co/genbio/33e5454… 2. Optimized CUDA Kernels for efficient foundation model computing, apply here: jobs.lever.co/genbio/824c21b… We’re also
Hongyi Wang
@HongyiWang10
Dec 11, 2023
Today, open source LLMs' pre-training details are unfortunately still hidden. We @llm360 seek to solve this by releasing ALL pre-training details, initiated with two 7B LLMs: Amber and CrystalCoder. We will keep pushing this effort forward. Come meet our team at #NeurIPS2023!
LLM360
@llm360
Dec 11, 2023
🚀 1/7 We are thrilled to launch LLM360 — pushing the frontier of open-source & transparent LLMs! Starting with Amber (7B) & CrystalCoder (7B), we are releasing brand new pre-trained LLMs with all training code, data, and up to 360 model checkpoints. 🔗 llm360.ai
00:00
Hongyi Wang
@HongyiWang10
Feb 26, 2024
I'm thrilled about this opportunity to visit @RutgersCS and talk about my work on LLM and distributed ML. Come to talk to me if you are on campus; see you tomorrow!
Rutgers Computer Science Department
@RutgersCS
Feb 26, 2024
Please join us tomorrow for our next Computer Science Colloquium. A Step Further Toward Scalable and Automatic Distributed Large Language Model Pre-training Tuesday, February 27, 2024, 10:30am - 12:00pm Speaker: @HongyiWang10 lnkd.in/eh_AwC7f
Hongyi Wang
@HongyiWang10
Nov 4, 2023
I'm thrilled to receive the @CPALconf Rising Star Award! I would like to thank the award committee for their consideration, and especially my advisors @DimitrisPapail and @ericxing for their unwavering support! I look forward to the conference!
Conference on Parsimony and Learning (CPAL)
@CPALconf
Nov 4, 2023
CONGRATS to our CPAL Rising Star Awardees!! An incredibly talented and impressive bunch 💯 See who they are and what they do: cpal.cc/rising_stars_a…
GIF
Hongyi Wang
@HongyiWang10
Jun 18, 2021
Successfully defended my Ph.D. dissertation and became a doctor! Many thanks to my amazing advisor @DimitrisPapail, and other great committee members @madsjw @thodrek @Kangwook_Lee and Shivaram Venkataraman! It is a great pleasure to have all of you.
Hongyi Wang
@HongyiWang10
Jun 17, 2023
Excited to present our tutorial on "ML Systems for Large Models & Federated Learning" @CVPR tomorrow at 8:30AM PST with Professor @qirong_ho and Professor @sam_hrvth @mbzuai!
Hongyi Wang
@HongyiWang10
Nov 30, 2023
Check out our new distributed training software, Redco, which will appear at @NeurIPSConf 2023! Redco simplifies training development on various GPU/TPU clusters. Our team has successfully used it to train multiple powerful LMs and LLMs. Life is short, so give Redco a try!
Bowen Tan
@BowenTan8
Nov 30, 2023
Thanks @hgpu for posting our work. So happy we're in the "MOST VIEWED PAPERS" on hgpu.org 🚀Find Redco at Mlsys workshop of #NeurIPS2023 -automatic distributed training with large models and complex algorithms. 📄 arxiv.org/abs/2310.16355 💻 github.com/tanyuqian/redco
Hongyi Wang
@HongyiWang10
Jan 16, 2024
FoE got in #ICLR2024! Come to talk to my colleagues in Austria about fusing expert models to boost performance effectively and efficiently. I probably won't be able to travel, unfortunately. But I'm still super proud of this paper!
Hongyi Wang
@HongyiWang10
Oct 4, 2023
1/ Our paper is out! We developed the Fusion of Experts (FoE) that merges expert model and LLM outputs for better performance across image & text classification, text summarization, MMLU, and automatic text evaluation. paper: arxiv.org/abs/2310.01542 Thread below.
Hongyi Wang
@HongyiWang10
Dec 19, 2024
Our team @llm360 is hiring at ALL LEVELS (interns and full-time employees)! Are you excited to dive into training, fine-tuning, and deploying LLMs? Apply NOW!
LLM360
@llm360
Dec 19, 2024
The LLM360 team is hiring! 🚀 We welcome researchers and ML engineers to join our team to push the frontier of AI research. You'll be deeply involved and gain firsthand experience in large model training. Join us in our mission of making AGI open to all. For more information,