Juncheng Yang Profile
Juncheng Yang

@1a1a11a

Followers
2K
Following
2K
Media
16
Statuses
679

Assistant Professor at Harvard University, Ph.D @ CMU, ex-Twitter, ex-Cloudflare.

Pittsburgh, PA
Joined September 2014
Don't wanna be here? Send us removal request.
@1a1a11a
Juncheng Yang
12 days
RT @JiaZhihao: 📢Exciting updates from #MLSys2025! All session recordings are now available and free to watch at We….
0
30
0
@1a1a11a
Juncheng Yang
1 month
RT @jeffra45: 🧵1/ New release from @Snowflake AI Research:. Shift Parallelism is a new LLM inference technique built on top of vLLM, relea….
0
18
0
@1a1a11a
Juncheng Yang
1 month
Super cool work speeding up your AI inference workloads and reducing cost from Snowflake AI Research!.
@AurickQ
Aurick Qiao
1 month
Excited to open-source Shift Parallelism, developed at @Snowflake AI Research for LLM inference!. With it, Arctic Inference + @vllm_project delivers:. 🚀3.4x faster e2e latency & 1.06x higher throughput.🚀1.7x faster generation & 2.25x lower response time.🚀16x higher throughput
Tweet media one
0
0
3
@1a1a11a
Juncheng Yang
1 month
RT @AurickQ: Very proud of our recent work at @Snowflake AI Research, which spans from the systems layer to the application layer. Check ou….
0
2
0
@1a1a11a
Juncheng Yang
2 months
RT @ACMSIGOPS: The Academic Pipeline Stall: Why Industry Must Stand for Academia, by Professor Vijay Janapa Reddi.
0
4
0
@1a1a11a
Juncheng Yang
2 months
Awesome work!.
@AurickQ
Aurick Qiao
2 months
Excited to share our work on Speculative Decoding @Snowflake AI Research!. 🚀 4x faster LLM inference for coding agents like OpenHands @allhands_ai. 💬 2.4x faster LLM inference for interactive chat . 💻 Open-source via Arctic Inference as a plugin for @vllm_project . 🧵
Tweet media one
0
0
1
@1a1a11a
Juncheng Yang
3 months
I got an email from Xfinity stating 99% reliability. My first reaction was, "Xfinity has so many outrages, how can you claim a high reliability?" Then I realized 99% means three days in a year without a network, so maybe they are correct. .
0
1
6
@1a1a11a
Juncheng Yang
3 months
We welcome everyone with system-building experience. this is a good opportunity for students and engineers to engage with the community!.
@giorgioercixu
Giorgio Erci Xu
3 months
Juncheng (@1a1a11a) and I are organizing Artifact Evaluation for #FAST26 and we are seeking members to join the commitee fo Spring and Fall cycles. Help promote reproducibility, and engage with cutting-edge storage research!. Please apply by June 2nd via 
0
0
4
@1a1a11a
Juncheng Yang
3 months
RT @minlanyu: Excited to co-lead with @Le_Xie_Energy a new short course at Harvard @hseas on May 21: Power Systems and AI: An Introduction.….
0
2
0
@1a1a11a
Juncheng Yang
3 months
I wondered why the Sieve cache website ( received thousands of visits last week. It turned out that some groups liked our website and copied it without changing the pageview tracking. I am happy that our work received so much attention!.
0
1
6
@1a1a11a
Juncheng Yang
4 months
just submitted a long paper review with more than 2000 words, I hope the authors will appreciate it.
0
0
17
@1a1a11a
Juncheng Yang
7 months
RT @du_yilun: We’re recruiting new tenure-track faculty in the Harvard Kempner Institute and computer science this year!. Apply here:. http….
0
3
0
@1a1a11a
Juncheng Yang
7 months
I am taking students in Fall 2025, if you are interested in storage systems, databases, and machine learning systems, please apply to the School of Engineering and Applied Science at Harvard University.
4
72
238
@1a1a11a
Juncheng Yang
7 months
RT @litian0331: I am taking new Ph.D. students from @UChicagoCS and @DSI_UChicago in the 2024-2025 cycle! If you are interested in distribu….
0
91
0
@1a1a11a
Juncheng Yang
9 months
RT @minlanyu: Harvard CS is hiring this year:.
0
18
0
@1a1a11a
Juncheng Yang
10 months
This is sad news.
@anandtech
AnandTech
10 months
After 27 years of providing in-depth coverage of the amazing world of PC and mobile hardware, AnandTech is saying farewell. We want to thank everyone from the AnandTech community for their support and passion for what we’ve done over the years.
0
0
4
@1a1a11a
Juncheng Yang
10 months
We need better ways to make this under control especially given that many are not reproducible.
@thegautamkamath
Gautam Kamath
10 months
H-index of top 20 publication venues in AI, April 2013 versus today. 2013: NeurIPS (37), ICML (65), ICLR (did not exist).2024: NeurIPS (337), ICML (268), ICLR (304). Please stop writing so many papers.
Tweet media one
Tweet media two
0
0
3
@1a1a11a
Juncheng Yang
10 months
Soon recruiters will use (if not already) bots to choose candidates, and we will have an army of bots fight another army of bots.
@rohanpaul_ai
Rohan Paul
10 months
Somebody uses an AI Bot to AUTOMATICALLY apply to 1000 JOBS in 24h and get 50 INTERVIEWS! 🤯. And the Bot's code is open-sourced in GitHub. 👨‍🔧. This tool automates your LinkedIn job search and application process. - Scans for openings matching your criteria,. - Applies to jobs
Tweet media one
0
0
2