Dan Fu

@realDanFu

Founding SVP of Research @togethercompute Assistant Professor @ucsd_cse Looking for talented kernel engineers and performance engineers!

danfu.org

Joined September 2019

246 Following

8,588 Followers

Replies Reposts Media

Pinned
Dan Fu
@realDanFu
Aug 19, 2024
Excited to share that I will be joining UCSD CSE as an assistant professor in January 2026! I'll be recruiting PhD students from the 2024 application pool - if you're interested in anything ML Sys/efficiency/etc please reach out & put my name on your application! Until then
50
Dan Fu
@realDanFu
Nov 4
Super excited to (officially) share Together Compiler with the world! This is our opinionated take at @togethercompute about how to scale inference in an increasingly polymorphic world (many models, archs, traffic, hardware). Deployed across our inference fleet today 🚀 1/
Simran Arora
@simran_s_arora
Nov 4
Introducing the Together Compiler: continual compilation for AI inference! I’m super proud to finally share this work from our research team! ✨ Today’s inference is “Polymorphic” or shape-shifting, it doesn’t sit still. On OpenRouter, 80% of traffic goes to models released in
GIF
6
Dan Fu
@realDanFu
Nov 4
PKB accepted to #neurips2026, go chat with @opengroundsFX and @simonguozirui in Sydney if you’re interested! Fun fact - arxiv actually rejected this paper, so my accept rate on arxiv this cycle is worse than NeurIPS :) Grateful for @askalphaxiv for an alternative!
Willy Chan
@opengroundsFX
Nov 4
Proud to announce that ParallelKernelBench (alphaxiv.org/abs/2606.paral…) has been accepted to NeurIPS 2026! 🎉 If you are interested in the future of multi-GPU kernels, come say hi! Huge thanks to my amazing collaborators @asplencmnt @simonguozirui @simran_s_arora @realDanFu
1
Dan Fu
@realDanFu
Oct 29
Introducing NyoomFloat12, a hardware-friendly lossless 12-bit quant from BF16 (accepted to #neurips2026)! In BF16 LLMs four of the bits that are the same for 99.7% of the weights. So you can compress BF16 -> 12-bit with minimal overhead. Check out @__sylvie_l__'s thread! 1/
Sylvie L
@__sylvie_l__
Oct 29
Lossless weight compression has saved LLMs GPU memory, but not time. Decoding costs more than it saves. NyoomFloat12 (NF12) is a lossless 12-bit BF16 format with decode cheap enough to speed up inference. Qwen3-32B: 1.23× faster token generation, up to 2.54× rollout throughput🧵
4
Dan Fu
@realDanFu
Oct 28
I had a great time at this launch event, thanks for having me @StepFun_ai!
StepFun
@StepFun_ai
Oct 28
Rain couldn't stop the builders. Great conversations at our Step 5 Preview Launch Night with @togethercompute! From long-horizon agents and tool-calling reliability to inference costs and what AI breakthroughs mean for the next wave of startups. Huge thanks to our panelist Dan
2

Dan Fu

Dan Fu’s posts