Hyunsung Kim

About

hyunsung.kim [at] kaist.ac.kr

Hi, I am a Sports Data Scientist who loves both mathematics and soccer. I am currently a Ph.D. student at Data Science & Artificial Intelligence Lab (DSAIL), Graduate School of Data Science at KAIST, advised by Prof. Chanyoung Park. Alongside my studies, I work as a part-time Data Scientist at Fitogether, an IT startup that provides a data analysis solution to sports clubs. Previously, I received a B.S. in Mathematics at POSTECH and an M.S. in Industrial Engineering at Seoul National University.

Research Interests

I'm interested in applying AI algorithms to sports analytics, where my specialty and passion intersect. More specifically, my current research topics are:

Sports data collection and preprocessing (MLSA'22, MLSA'25)
Multi-agent trajectory prediction and imputation (KDD'23, PKDD'25)
Player fatigue and injury prediction (Access'22)

Playing style representation and similar player retrieval (MLSA'21)
Tactical pattern detection and classification (KDD'22, IT4PSS'24)
Performance evaluation and optimization (MLSA'23)

The above topics can boil down to modeling multi-agent trajectories. Ideally, I aim to build a foundation model for soccer that can solve the above problems through downstream tasks.

News

Jul 2025: I gave a presentation about wearable-optical fusion tracking in sports at NU Sports ML Workshop 2025, organized by Sports Behavior Group at Nagoya University.
Jul 2025: One paper (W5) got accepted to MLSA 2025. Glad to present one paper (C3) at the main conference and one paper (W5) at a workshop of ECML PKDD 2025!
May 2025: One paper (C3) co-authored with Han-Jun Choi got accepted to ECML PKDD 2025.
May 2025: Invited by Yonsei Sports Analytics Lab, I gave a presentation about my research at Yonsei University.
Dec 2024: Invited by Prof. Sang-Ki Ko, I gave a presentation about my research at the Department of AI, University of Seoul.
Sep 2024: I joined DSAIL at KAIST and began to pursue a Ph.D. degree!
Jun 2024: One paper (W4) got accepted to IT4PSS 2024.
Mar 2024: I'll participate IT4PSS 2024 workshop held in conjunction with IJCAI 2024 as an organizing committee.
Feb 2024: I was invited to Dagstuhl Seminar 24081: Computational Approaches to Strategy and Tactics in Sports and spent an awesome week meeting leaders of worldwide sports analytics community.
Dec 2023: Invited by Marc Schmid, I gave an online presentation about my SoccerCPD paper (C1) to members of Chair of Performance Analysis and Sports Informatics at TU Munich.
Jul 2023: One paper (W3) led by Pegah Rahimian got accepted to MLSA 2023. Special thanks to Pegah for inviting me an interesting study!
May 2023: One paper (C2) got accepted to KDD 2023. Together with the paper (C1) published in KDD 2022, I get papers at KDD for two years in a row!
Oct 2022: Fitogether has been selected as a FIFA Preferred Provider (FPP). So far, there are only two FPPs (Fitogether and Kinexon) around the world and we are the first GPS tracking data provider collaborating with FIFA as an FPP.
Aug 2022: One paper (J1) led by Jeongbin Kim got accepted to IEEE Access.
Aug 2022: I have launched Soccer Data Analytics with Python course in Class101. To my knowledge, this is the first Korean online course on soccer data analytics.
Jul 2022: One paper (W2) got accepted to MLSA 2022.
Jul 2022: My interview with SPOTV News, a Korean online press, has been published in Naver Sports. I talked about the growing share of data-driven decision-making in the soccer ecosystem.
May 2022: One paper (C1) got accepted to KDD 2022. This will be my first publication in a top AI/DS conference!
Dec 2021: I was invited as a guest speaker to the 16th K League Football Industry Academy, and made a presentation about my career and research work. Thank K League as a fan and an academy alumnus for inviting me again!

Jul 2021: One paper (W1) got accepted to MLSA 2021. This will be my first presentation in the sports analytics academia!