English
Hong Jin Kang
347 posts
Hong Jin Kang
@kanghj91
Lecturer (equivalent to assistant professor in the US system) at University of Sydney. Learn more at https://t.co/XsvRkQfH0o!
Katılım Mayıs 2013
489 Takip Edilen181 Takipçiler
Happy to share that i have joined @GoogleAI @GoogleLabs as a scientist to build the next-generation coding agents.
The next-generation coding agent should be the one that can autonomously do tasks without supervision, self-improving, and can be seen as a stepping stone towards SuperIntelligence.
My approach stays the same: fast iteration between research ideas and user-facing products. Build it, ship it, learn from real users, and innovate from there. Believe this is the only way to thrive in the AI race and @GoogleLabs could be the right place to do this.
English
Thrilled to announce that I'll be joining UIUC CS @siebelschool as an Assistant Professor in Spring 2026!
📢 I’m looking for Fall '26 PhD students who are interested in the intersection of Software Engineering and AI, especially in LLM4Code and Code Agents. Please drop me an email if you are interested in working with me.
English
English
Great time presenting 2 papers @IEEEICSME (Auckland 🇳🇿) on Just-In-Time #Vulnerability Prediction, enabling shift-left security.
🧪 Realistic evaluations cause big model perf. drops + lessons learned: arxiv.org/abs/2507.10729
🔧 VulGuard tool for JIT-VP: arxiv.org/abs/2507.16685
English
lionguard 2 has been accepted at @emnlpmeeting system demonstration track! 🥳🥳🥳
joint work with leanne tan, ge ziyu (zoey), and @sroylee
Gabriel Chua@gabrielchua
today, we’re releasing 𝐥𝐢𝐨𝐧𝐠𝐮𝐚𝐫𝐝 2 🦁✌️ our latest llm guardrail covering english/singlish, chinese, malay, and tamil we're releasing the model weights, parts of our training data and a detailed technical report 🧵 @SRoyLee @GovTechSG @sutdsg
English
I am happy to share that I passed the dissertation defense at SMU today! I am now officially Dr. Tu! (Or Dr. Two🤣, as I finally completed the Dual-degree PhD program 🎓🎓 in 6 years!)
Special thanks to all committee members: Profs. Lingxiao, Xuhua, @davidlo2015, and @mboehme_!
English
English
🚨 Our paper is early accepted at @IEEEICSME 2025! 🎉 (6% rate)
Takeaways:
📉 JIT #Vulnerability prediction drops up to 98% in realistic settings due to data imbalance
⚖️Custom solutions are needed to ensure the practicality of JIT-VP
📄Preprint coming soon!🇦🇺🤝🇻🇳@UniofAdelaide
English
📣 Just got awarded two grants (Co-PIs): $4.9M for “LLM Architecture-Language Nexus: A Unified Analysis Framework” & $3.4M for “Heterogeneous Preference Learning for Value-Aligned Large Language Models in Multicultural Societies.” Recruiting Research Assistants & Fellows—PM me!
English
Thank you very much, @msrconf, award committee, nominator, and endorser for the help, support, and recognition, which are very much appreciated :) Thank you very much to all my mentors, students, and collaborators for the help and support :)
MSR 2025@msrconf
We are thrilled to announce that #MSR2025 Foundational Contribution Award 🏆💫goes to David Lo @davidlo2015 for pioneering, influential, and lasting contributions to transforming bug and test data into insights and automation that improve software quality and productivity.
English
@asankhaya Very cool!
We are working on something similar with SMU. It’ll be interesting to compare results once we complete our evaluations!
English
🔐 Just released AutoGrep! AI-powered generation of Semgrep/OpenGrep security rules from CVE patches.
Processed 40K patches → 645 high-quality rules across 20 languages.
No more manual rule curation!
Code: github.com/lambdasec/auto…
Paper: lambdasec.github.io/AutoGrep-Autom…
#AppSec #SAST
English
🎉 Excited to share that my first paper at #ICLR @iclr_conf has been accepted! Big thanks to my awesome colleague @marcel and the fantastic SoftSec group #MPI-SP @maxplanckpress. Marcel has shared a great post with a brief explanation of our work—check it out! 🚀
Marcel Böhme👨🔬@mboehme_
Just got our first @ICLR_conf paper accepted! 🥳🥳 It shines light on a beautiful riddle! Suppose, you're drawing balls from an urn with an unknown distribution of colors. What is the proportion of balls with colors, you've never seen? Led by Seongmin (@nim_gnoes_eel).
English
"Dear Dr. Menzies
It is my privilege to confirm your elevation to ACM Fellow. Fellow is ACM's highest membership grade. The number of Fellows is limited and the criteria are very strict. Congratulations on this very significant honor."
I have so many people to thank...
English
Hong Jin Kang retweetledi
This year, @rodrigoospinola & I are pleased to introduce the new AI Models and Data Track at @ConfEASE (CORE A).
Please consider submitting your exciting work, including datasets, tools, and AI innovations for #SoftwareEngineering.
👉conf.researchr.org/track/ease-202…
📅 Deadline: Mar 16
English
🌟 🌟 Excited to share that our RevFinder paper published at SANER2015 has won the SANER2025 Most Influential Paper (MIP) award. 🌟 🌟
It was one of the first to initiate an important step in paving the "Automated Code Reviews" which has been researched extensively nowadays.
English
Hong Jin Kang retweetledi
I have background in mining software repositories. Until I see a peer reviewed study, this is extremely suspect:
Measuring productivity by looking at code output is... bad
Yegor Denisov-Blanch@yegordb
Another way to look at this is counting code commits. While this is a flawed way to measure productivity, it reveals inactivity: ~58% make <3 commits/month, aligning with our metric. The other 42% make trivial changes, like editing one line or character--pretending to work.
English
Look forward to an exciting keynote!
sea4dq@sea4dq
We are thrilled to welcome Qinghua Lu @qinghualu as a keynote speaker at the #SEA4DQ workshop! As the Principal Research Scientist at CSIRO's Data61 @Data61news, she'll be discussing Responsible AI engineering from a data perspective. Don't miss her insights on July 15 in Brazil!
English