Minchul Kang

Papers

* first author

1. 1

   Parallelism Strategy Chaining for Fast Training Convergence

   Minchul Kang, Changyong Shin, Younghun Go, Hyunho Lee, Jinwoo Jeong, Chuck Yoo, Gyeongsik Yang

   EMNLP'26 Main (acceptance rate: 15.4%)
2. 2

   GeoMesh: Workload-Balanced and Sign-Compressed Geo-Distributed LLM Training

   Changyong Shin, Jaerim Park, Minchul Kang, Younghun Go, Zhixiong Niu, Yongqiang Xiong, Gyeongsik Yang, Chuck Yoo

   EMNLP'26 Findings
3. 3

   Unified KV Pooling to Accelerate Long-Context LLM Serving

   Minchul Kang, Changyong Shin, Jinwoo Jeong, Jaerim Park, Woohyun Kim, Bonyul Gu, Dongwoo Kang, Gyeongsik Yang, Chuck Yoo

   Preprint'26
4. 4

   Making Sense of Job Preemption for Distributed Deep Learning Acceleration

   Younghun Go, Changyong Shin, Minchul Kang, Jae Hyun Hwang, Chuck Yoo, Gyeongsik Yang

   DAC'26
5. 5

   GPU Memory Prediction for Multimodal Model Training

   Jinwoo Jeong, Minchul Kang (co-first author), Younghun Go, Changyong Shin, Hyunho Lee, Junho Yoon, Gyeongsik Yang, Chuck Yoo

   SAA '25 @ SOSP 2025
6. 6

   Training Time Prediction for Mixed Precision-based Distributed Training

   Minchul Kang

   Preprint'26

Open Source

• pyftrace — low-overhead Python function tracer

  Maintainer · 200+  on GitHub

• csnote — computer science keyword notes

  Maintainer · 295+  on GitHub

• uftrace — function graph tracer for C/C++/Rust/Python

  Contributor · 3.4k  on GitHub · Top-5 contributor (50+ commits)

Books

• Computer Architecture and Operating Systems for Self-Study

  Selected as "Computer Science Book of the Year" (Kyobo, 2022). Top-20 IT bestseller in Korea for 95 weeks.

• Networking for Self-Study

  Top-20 IT bestseller in Korea for 14 weeks.

• Git & GitHub for Everyone

  Selected as "Sejong Excellent Academic Book" (KPIPA, 2023).

• This Is Computer Science

  Top-100 IT bestseller in Korea for 18 weeks.