📃 Curriculum Vitae

:googlescholar: Google Scholar

:linkedin: Linkedin

image.png

Hi, My name is Junsu Kim.

I am on a journey searching for ideas and people that can make a fundamental **change to the world.

Currently my primary goal is understanding intelligence, especially thinking-the state of the art method humanity has ever had to solve a new problem. As a student researcher in Google DeepMind,  I am focusing on exploring thinking.

Furthermore, I am also interested in the anthropologic and economic consequences of intelligence.

My journey stems from **Ryu Optimization Group,** where I first dived in the field of artificial intelligence and learned how to reasearch. Before that, I enjoyed my time preparing for IMO, where I first learned how to think*.*

If you'd like to chat, I'm hosting a **30-minute Chat** every day. Feel free to connect-I love talking with new people!

Publications

image.png

LoRA Training Provably Converges to a Low-Rank Global Minimum or It Fails Loudly (But it Probably Won’t Fail)

Junsu Kim, Jaeyeon Kim, Ernest K. Ryu

International Conference of Machine Learning (Oral Presentation), 2025

image.png

Clip-Low Increases Entropy and Clip-High Decreases Entropy in Reinforcement Learning of Large Language Models

Jaesung R. Park, Junsu Kim, Gyeongman Kim, Jinyoung Jo, Sean Choi, Jaewoong Cho, Ernest K. Ryu

arXiv Manuscript, 2025.

image.png

Towards Robust Mathematical Reasoning

Thang Luong◊, Dawsen Hwang*, Hoang Nguyen*†, Golnaz Ghiasi*, Yuri Chervonyi*, Insuk Seo*, Junsu Kim*, Garrett Bingham, Jonathan Lee, Swaroop Mishra†, Alex Zhai, Clara Huiyi Hu, Henryk Michalewski, Jimin Kim†, Jeonghyun Ahn†, Junhwi Bae†, Xingyou Song, Trieu H. Trinh, Quoc V. Le, Junehyuk Jung◊

Empirical Methods in Natural Language Processing, 2025

Meetings

https://calendar.app.google/ZdgpUiUjm2D3ToEo6