Linhong Zhu

Engineering leader at Meta focused on agentic AI, recommender systems, and LLM-native products

[Linhong Zhu @smileofice]

Excited to see Muse Code released in beta! Fortunate to be part of Meta’s Coding Acceleration V-Team and to have supported coding-model evaluation. Proud to have played a small part—huge congratulations to the teams who brought it to life!

[Alexandr Wang @alexandr_wang]

good iteration for muse spark 1.2

onto bigger and better things 🍉

[Mark Zuckerberg @finkd]

Releasing Muse Code in beta today. It's a terminal coding agent that takes on complete software engineering tasks across large repos: planning changes, writing code, validating the results. Powered by Muse Spark 1.2, a coding-focused model update.

[Linhong Zhu @smileofice]

Excited to share two works from our team on LLM-powered recommendation:
• Multimodal LLM framework for large-scale recommendation — #SIGIR2026: https://t.co/oGwZlsMfvk
• Shape Your Feed, an agentic conversational recommender — #RecSys2026. Camera-ready PDF coming soon.

[Linhong Zhu @smileofice]

Back on X after almost 10 years away.

A lot has changed: especially in AI.
Feels like one of the few places where ideas, research, products, and culture now collide in real time.

Looking forward to learning, sharing, and reconnecting with people 😀

[AI Research at ISI @ISI_AI]

Congratulations to @AI_ISI’s @smileofice Majid @szeke @aram_galstyan @caknoblock for winning the @ISWC2016 Best Paper Award! @USCViterbi https://t.co/IRtrtOLJ0m

[Linhong Zhu @smileofice]

The first best paper ! Oh Yeah! https://t.co/SxIoSlNIow

[Pedro Szekely @szeke]

Won the best paper award at #iswc2016 https://t.co/AWMFaPuFN0 #digusc

[Aram Galstyan @aram_galstyan]

Our work on Twitter bot detection mentioned on Mashable https://t.co/t5ZmYtIF95

[Pedro Szekely @szeke]

Unsupervised Entity Resolution via Graph Summarization nominated for best paper, Thu 11:30am, #digusc #iswc2016

[Aram Galstyan @aram_galstyan]

Unsupervised Entity Resolution via Graph Summarization, our @ISWC2016 paper w/ @smileofice Majid @szeke @caknoblock https://t.co/f9eReZ3mHu

[Linhong Zhu @smileofice]

#graphembedding Reading 3: https://t.co/va5bhy6tIF (use negative sampling as objective to embed nodes based on links and attributes)

[🇺🇦 Kristina Lerman 🇺🇦 @KristinaLerman]

Biz Insider reports on our participation in the DARPA Twitter Bot Challenge - https://t.co/OkgrHZ7rF6 via @bi_enterprise

[Linhong Zhu @smileofice]

#graphembedding reading 2: Semi-supervised Learning with
Deep Generative Models: https://t.co/MpWPDiTEZY

[Linhong Zhu @smileofice]

Summary of the paper: learn the distributed representation of nodes in networks via considering nodes as words, random walks as sentences.

[Linhong Zhu @smileofice]

Paper on Today's reading list: DeepWalk: Online Learning of Social Representations (https://t.co/VciBA6CHvT) #graphembedding

[🇺🇦 Kristina Lerman 🇺🇦 @KristinaLerman]

Holiday lunch with my students at a local tourist trap.

[Linhong Zhu @smileofice]

Make your name visible to attract new followers!!! see what we found here: http://t.co/FT3iRp9Geo

[Linhong Zhu @smileofice]

Want to test whether high user activity will attract new followers or lose followers. So I will start tweeting daily and see how....

[Linhong Zhu @smileofice]

@lionmanu sounds great:)

[🇺🇦 Kristina Lerman 🇺🇦 @KristinaLerman]

The Curses of Heterogeneity in Big Data - my new blog post http://t.co/7R4Q2GFhTP

[Shumo Chu @shumochu]

三聯最近出了一個很有格調的雜誌，中文名叫新知（Wissen），主要定位于軟硬兼備的科普，雜誌的motto就逼格爆棚：Wir müssen wissen, wir werden wissen. (we must know, we will know. )