Currently a Postdoctoral Researcher at
Embedded Intelligent Systems (EIS) Lab, Yonsei University (PI: Prof. JeongGil Ko)
Human-Robot Interaction (HRI) Lab, KAIST (PI: Prof. Ki-Uk Kyung)
Previously a Visiting Researcher at
Keio University @ Prof. Jin Nakazawa and Prof. Tadashi Okoshi's Lab
University of Michigan @ Prof. Pei Zhang's Lab
i.am.sungmin.lee [at] yonsei.ac.kr
i.am.sungmin.lee [at] gmail.com
Sep, 2026 — I am joining HRI Lab @ KAIST as a Postdoctoral Researcher while continuing my affiliation with EIS Lab @ Yonsei, funded by the INNOCORE program, to work on Efficient AI for Real-time Robotics!
June, 2026 — I've succesfully defended my Ph.D. thesis. Thanks to everyone, including my advisor Prof. Jeonggil Ko, commitee members, and all my colleages.
April, 2026 — Our work JARVIS for HVAC: LLM-based Question-Answer Framework for Sensor-driven HVAC System Interaction has been accepted to ACM UbiComp/IMWUT 2026.
Feb, 2026 — A total of four papers, including Spatial Discretization for Fine-Grain Zone Checks with STARKs (Spotlight), have been accepted to the ACM Web Conference 2026 Workshop on Zero-Knowledge Proofs and Blockchain for Web 4.0 (ZABAPAD 2026).
Dec, 2025 — Our work Foreground Graphics-Aware Runtime DNN Scheduling on Mobile GPUs, conducted with Minju and Jaeho, has been accepted to the interactive session of IEEE PerCom 2026!
Oct, 2025 — I am attending IEEE ISMAR 2025 in Daejeon, Korea to present my work, Hierarchical Localization and Tracking for In-Vehicle Mixed Reality.
Jan, 2025 — I am visiting Professor Pei's lab at University of Michigan as a visiting researcher from Feb 2025 to July 2025!
Dec, 2024 — I am visiting JNLab at Keio University as a visiting researcher during Dec 2024!
Dec, 2024 — My work with Kichang on designing a light-weight classifier using gZip for extremely constraint platforms has been accepted to ICT Express!
Jun, 2024 — My work with Kichang on designing an ultra-lightweight time-series classification model with gZip compression received the Best Poster Runner-Up Award from ACM MobiSys 2024!
Sungmin received his Ph.D. in Engineering from Yonsei University under the supervision of Prof. JeongGil Ko.
He is currently a postdoctoral researcher at the Human-Robot Interaction (HRI) Lab at KAIST and the Embedded Intelligent Systems (EIS) Lab at Yonsei University, working on efficient AI for real-time robotics.
He was a visiting researcher at JN Lab, Keio University, in 2024 and Prof. Pei Zhang’s Lab at the University of Michigan in 2025.
His research focuses on resource-efficient time-series sensing intelligence and on-device AI, with recent interests extending to intelligent 6G systems and resource-efficient AI for robotics.
MS&Ph.D., 2021.03-2026.08, School of Integrated Technology, Yonsei University, Seoul, Korea
Advisor: Prof. JeongGil Ko
Dissertation: Resource-Efficient Time-Series Inference under Post-Deployment Dynamics across the Sensor-to-Service Computing Stack
B.Eng., 2018.03-2020.02, School of Integrated Technology, Yonsei University, Seoul, Korea
Track (Major/Minor) : Computer Science / Robotics
Under Review/Revision
Conferences/Workshops
[C5, 2026] Ruiqing Wang, Jiale Zhang, Sungmin Lee, Jeremy Miles, Gary Rohrer, Yiwen Dong. "Behav2Need: Ethology-Driven Animal Need Understanding via Multimodal Sensing on Smart Farms", ACM BuildSys 2026 [pdf]
[C4, 2026] Sungmin Lee, Joonhee Lee, Minju Kang, Seungyong Lee, Dongju Kim, Jingi Hong, Jun Shin, Pei Zhang, JeongGil Ko. "JARVIS for HVAC: LLM-based Question-Answer Framework for Sensor-driven HVAC System Interaction", ACM International Joint Conference on Pervasive and Ubiquitous Computing (ACM UbiComp) 2026 [pdf]
[W3, 2026] Kichang Lee, Sungmin Lee, Jaeho Jin, JeongGil Ko. "Verifiable Dropout: Turning Randomness into a Verifiable Claim", ACM Web Conference 2026 Workshop on Zero-knowledge proof And Blockchain for WEB 4.0 (ZABAPAD) 2026, Dubai, UAE. [pdf]
[W2, 2026] Gyeongmin Han, Sungmin Lee, Joonhee Lee, Kichang Lee, JeongGil Ko. "Statistical Analysis of 12V Power Load Uncertainty Using Conditional Probability for SDV Energy Management", ACM Web Conference 2026 Workshop on Zero-knowledge proof And Blockchain for WEB 4.0 (ZABAPAD) 2026, Dubai, UAE. [pdf]
[W1, 2026] Sungmin Lee, Kichang Lee, Gyeongmin Han, JeongGil Ko. "Spatial Discretization for Fine-Grain Zone Checks with STARKs", ACM Web Conference 2026 Workshop on Zero-knowledge proof And Blockchain for WEB 4.0 (ZABAPAD) 2026, Dubai, UAE. (Spotlight) [pdf]
[C3, 2026] Minju Kang, Jaeho Jin, Sungmin Lee, JeongGil Ko. "Foreground Graphics-Aware Runtime DNN Scheduling on Mobile GPUs", IEEE International Conference on Pervasive Computing and Communications (IEEE PerCom) 2026, Pisa, Italy. (Acceptance Rate: 15.62%; NRF-listed Top CS Conference) [pdf]
[C2, 2024] Yujin Shin*, Kichang Lee*, Sungmin Lee, You Rim Choi, Hyung-Sin Kim, JeongGil Ko. "Effective Heterogeneous Federated Learning via Efficient Hypernetwork-based Weight Generation", ACM Conference on Embedded Networked Sensor Systems (ACM SenSys) 2024, Hangzhou, China. (Acceptance Rate: 18.5%; NRF-listed Top CS Conference) [pdf]
[C1, 2023] JaeYeon Park, Kichang Lee, Sungmin Lee, Mi Zhang, JeongGil Ko. "AttFL: A Personalized Federated Learning Framework for Time-series Mobile and Embedded Sensor Data Processing", ACM International Joint Conference on Pervasive and Ubiquitous Computing (ACM UbiComp) 2023. Cancun, Mexico. (NRF-listed Top CS Conference) [pdf]
Journals
[J5, 2026] Sungmin Lee, Minju Kang, Joonhee Lee, Seungyong Lee, Dongju Kim, Jingi Hong, Jun Shin, Pei Zhang, JeongGil Ko. "JARVIS for HVAC: LLM-based Question-Answer Framework for Sensor-driven HVAC System Interaction", Proceedings of the ACM (PACM) on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT) 2026 [pdf]
[J4, 2024] Sungmin Lee*, Kichang Lee*, JaeYeon Park, JeongGil Ko. "GMT: Gzip-based Memory-efficient Time-series Classification", ICT Express 2024 [pdf]
[J3, 2023] JaeYeon Park, Kichang Lee, Sungmin Lee, Mi Zhang, JeongGil Ko. "AttFL: A Personalized Federated Learning Framework for Time-series Mobile and Embedded Sensor Data Processing", Proceedings of the ACM (PACM) on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT) 2023 [pdf]
[J2, 2023] Sungmin Lee*, Hyungi Kim*, Kyung Won Lee, Eun Su Kim, Hyung-Mo Kim, Hyungsoon Im, Hyun C Yoon, JeongGil Ko, Eunha Kim. "Indolizine-based fluorescent compounds array for noninvasive monitoring of glucose in bio-fluids using on-device machine learning", Dyes and Pigments 2023 (JCR IF Top 10%) [pdf]
[J1, 2021] Sungmin Lee*, Hyungi Kim*, Jun Sik Min, Eunsu Kim, Junwon Choi, JeongGil Ko, Eunha Kim. "Fluorescent sensor array for high-precision pH classification with machine learning-supported mobile devices". Dyes And Pigments 2021 (JCR IF Top 5%) [pdf]
Posters/Demos
[P2, 2025] Daisuke Goto, Sungmin Lee*, Akira Tsuge*, Shuri Marui, Soko Aoki, JeongGil Ko, Tadashi Okoshi. "Hierarchical Localization and Tracking for In-Vehicle Mixed Reality". IEEE ISMAR 2025, Daejeon, Korea
[P1, 2024] Sungmin Lee*, Kichang Lee*, JeongGil Ko. "A Memory Efficient Parameter-free Time-series Classification via gzip". ACM MobiSys 2024, Tokyo, Japan. (Best Poster Runner-up Award) [pdf]
* Equal contributions
Spotlight Award, ACM Web Conference, ZABAPAD workshop 2026
Visiting Researcher, University of Michigan — 2025 (Full Funding)
Visiting Researcher, Keio University — Dec, 2024 (Full Funding)
Best Poster Runner-up Award, ACM Mobisys 2024
Best Project Award, Inter-university Alliance for Furture vehicular Technology(I4FT), Yonsei University — Dec, 2023
Best Project Award, Inter-university Alliance for Furture vehicular Technology(I4FT), Yonsei University — Jun, 2023
Full Scholarship and One Year Early Graduation, School of Integrated Technology, Yonsei University — 2018-2020
Duration: 2025 - 2026
Tasks: Data processing, inspection, tool developement
Partners:
University of Michigan, US
University of Illinois Urbana-Champaign, US
Duration: 2024.04 - 2025.04
Role: Lead Researcher
Partners:
LG Electronics (Home Appliances & Air Solutions), Korea
Experiences:
Time-series DB
LLM engineering
Query autogeneration
Duration: 2024.02 - 2024.06
Role: Lead Developer
Partners:
College of Medicine, Yonsei University (Prof. Kihyuck Kwak)
Experiences:
Image processing and statical analysis
UI/UX design
Windows GUI programming
Duration: 2023.05 - 2023.12
Role: Lead Researcher
Partners:
LG Electronics (Home Appliances & Air Solutions), Korea
Experiences:
Self-supervised clustering on unlabeled time-series data
HVAC zone clustering
Duration: 2021.03 - 2024.02
Role: Researcher
Partners:
Korea, Republic of
KETI, SNU, Security Platform
Sweden
RISE, Husqvarna, LumenRadio, ASSA ABLOY
Romania
BEIA
Experiences:
Building AI model for Flooding Attack Detection in RPL networks
Implementing FDTLS with Contiki-NG
Factory-scale real-world sensor data analysis
Duration: 2021.03 - 2023.02
Role: Researcher
Partners:
EPITONE, Korea
Experiences:
3D Pointcloud Compression
Visual-based Pose refinement
Software Project (Undergraduate), Yonsei Univ— Fall 2023, 2024
Advanced Algorithms (Undergraduate), Yonsei Univ— Fall 2021
Advanced Programming (Undergraduate), Yonsei Univ — Spring 2021
Student Volunteer Chair
ACM MobiSys 2024
Reviewer
IEEE TMC 2025
ACM Transactions on Internet of Things 2026
ACM CHI 2025
ACM IMWUT 2023, 2026
[A] 우리가 몰랐던 일상 속 인공지능 (Artificial Intelligence in Our Daily Lives), Center for Digital Social Science, Yonsei University — August, 2026 [link]
[PA] Intermittent Channel Estimation via Low-Cost Channel Dynamics Monitoring, 한국통신학회 — 2026
[T] Conversation with France President Emmanuel Macron, Yonsei University — April, 2026 [Korean article]
[T] Invited talk, KAIST, Korea — Oct, 2025
[T] Invited talk, Dankook University, Korea — Aug, 2025
[T] 7th A3 Foresight Workshop, Kyoto, Japan. — Feb, 2024
[P] KICS Winter conference, — Feb, 2024
[PA] 기계학습 기반 DIS Flooding Attack 감지 시스템 구현, 반도체공학회 — 2023
[T, D] STACK final project review, Sweden — Nov, 2023
[T, P] Inter-university Alliance for Furture vehicular Technology(I4FT), Yonsei University, Korea — Dec, 2023
[T, P] Inter-university Alliance for Furture vehicular Technology(I4FT), Yonsei University, Korea — Jun, 2023
Contiki-NG, Cooja & Multi-trace
NVIDIA Sionna-RT