KAIST Embedded AI Lab
(임베디드 인공지능 연구실)
School of Electrical Engineering (EE) · Department of AI Systems (AIS)Korea Advanced Institute of Science and Technology (KAIST)
Embedded Files
Building Embedded Physical AI Systems
Building Embedded Physical AI Systems
We build embedded physical AI systems that can sense, learn, reason, plan, adapt, and act across cyber and physical environments under real-world constraints on computation, memory, energy, and connectivity.
Our research spans both AI models and systems. Through full-stack model–system co-design and co-optimization, from foundation models and learning algorithms to runtimes, compilers, and hardware, we make advanced AI deployable, adaptive, efficient, and dependable on real-world devices.
Research Areas
AI Systems, Operating Systems, Runtimes, and Compilers: System software for efficient, portable, and dependable AI execution across heterogeneous embedded hardware.
Device-Native Foundation Models and Agents: Multimodal foundation models and autonomous agents co-designed from the ground up with real-world device constraints.
On-Device Learning and Continual Adaptation: Learning, personalization, and continual adaptation directly on resource-constrained devices.
Announcement
Announcement
We are seeking outstanding postdoctoral researchers to join our lab in 2026 and 2027. If you are interested, please submit your application through the following form: https://forms.gle/wHdMqzuxopsnZrRJ6
We are looking for new motivated members (Intern/MS/PhD) for 2026/2027 who will be part of an exciting journey at the School of Electrical Engineering (EE) and Department of AI Systems (AIS) at KAIST. Please take a look at our research statement and information for prospective students.
News
News
Aug 2026: “Intelligent Onboard Crater Detection for Lunar Exploration Systems Using Knowledge Distillation” received IAA SpaceAI 2026 Best Student Paper Award.
Aug 2026: Yuhyun Lee and Minseong Kim joined our lab. Welcome!
Jul 2026: Hyunsun Hong, Sungjeh Oh, and Sangin Mun joined our lab. Welcome!
Apr 2026: Selected for the 2026 Google Research Award (Cloud Research Credits)
Mar 2026: Wonheng Lee and Dongseok Ji joined our lab. Welcome!
Jan 2026: "Enabling Fine-Tuning of Direct Feedback Alignment via Feedback-Weight Matching" is accepted to ICLR 2026.
Jan 2026: "ECAI: Efficient Convolution Activation Inversion for Constant-Memory Convolutional Neural Networks Training" is accepted to AISTATS 2026.
Dec 2025: "Beyond Communication: On-Modem DNN Inference and Training for Home Appliances" is accepted (with shepherding) to PerCom 2026.
Nov 2025: "Mobile-Oriented Video Diffusion: Enabling Text-to-Video Generation on Mobile Devices Without Retraining, Compression, or Pruning" is accepted to WACV 2026.
Oct 2025: Isu Jeong has been selected as a finalist for the Qualcomm Innovation Fellowship Korea 2025 (QIFK 2025).
Sep 2025: "AliO: Output Alignment Matters in Long-Term Time Series Forecasting" is accepted to NeurIPS 2025.
Sep 2025: "Designing Extremely Memory-Efficient CNNs for On-device Vision and Audio Tasks" is accepted to IJCV (International Journal of Computer Vision).
Aug 2025: Jaeho Kim and Yunseok Lee successfully graduated!
Aug 2025: “Smart ECU: Scalable On-Vehicle Deployment of Drivetrain Fault Classification Systems for Commercial Electric Vehicles” is accepted to CIKM 2025.
Jul 2025: “Position: The AI Conference Peer Review Crisis Demands Author Feedback and Reviewer Rewards” got ICML 2025 Outstanding Position Paper Award. [award][link]
May 2025: Seulki Lee is honored with the Distinguished TPC Member Award at SenSys 2025.
May 2025: “TransPL: Pseudo-Labeling via Code Transitions for Time Series Adaptation” is accepted to ICML 2025.
May 2025: “Position: Towards a Sustainable AI Conference Peer Review System with Author Feedback and Reviewer Rewards” is accepted to ICML 2025 [Oral].
Mar 2025: “Bayesian Code Diffusion for Efficient Automatic Deep Learning Program Optimization” is accepted to OSDI 2025.
Jan 2025: “PPT: Patch Order Do Matters In Time Series Pretext Task” is accepted to ICLR 2025.
Jan 2025: Jaeyoung Lim joined our lab. Welcome!
Dec 2024: “SMMF: Square-Matricized Momentum Factorization for Memory-Efficient Optimization” is accepted to AAAI 2025.
Dec 2024: “Designing Extremely Memory-Efficient CNNs for On-device Vision Tasks (ACCV 2024)” got Best Application Paper Award.
Dec 2024: Jaeyun Shim joined our lab. Welcome!
Nov 2024: Jaeho Kim receives 3rd place award in K-Data Science Hackathon 2024.
Sep 2024: “Designing Extremely Memory-Efficient CNNs for On-device Vision Tasks” is accepted to ACCV 2024 [Oral].
Sep 2024: Jaeho Kim receives Best Poster Award (CAFO) in 2024 UNIST AI Technology Open Workshop.
Aug 2024: Jiwoon Chang and Yeojin Lee joined our lab. Welcome!
May 2024: Our paper, "CAFO: Feature-Centric Explanation on Time Series Classification", is accepted to KDD 2024 [Oral].
Apr 2024: Changmuk Lee joined our lab. Welcome!
Mar 2024: Bosung Kim receives 2nd Place Award from 2024 Hyundai Mobis Edge AI Research Contest (2024 현대모비스 Edge AI 산학협력 연구공모).
Mar 2024: Jaeho Kim receives 4th Place Award from 2024 Hyundai Mobis Edge AI Research Contest (2024 현대모비스 Edge AI 산학협력 연구공모).
MAR 2024: Seokhwa Hong joined our lab. Welcome!
Jan 2024: Our paper, "On-Demand Federated Learning for Arbitrary Target Class Distributions", is accepted to AISTATS 2024.
Sep 2023: Our paper, "Softmax Output Approximation for Activation Memory-Efficient Training of Attention-based Networks", is accepted to NeurIPS 2023.
Sep 2023: Our paper, "On-NAS: On-Device Neural Architecture Search on Memory-Constrained Intelligent Embedded Systems", is accepted to SenSys 2023.
Aug 2023: Yunseok Lee joined our lab. Welcome!
Jun 2023: Yoel Park and Kyu Hwan Lee joined our lab. Welcome!
May 2023: Our paper, "MicroDeblur: Image Motion Deblurring on Microcontroller-based Vision Systems", got Best Paper Runner-Up Award at IPSN 2023.
Apr 2023: Our paper, "Multitask Deep Learning for Human Activity, Speed, and Body Weight Estimation using Commercial Smart Insoles" is accepted by IEEE Internet of Things Journal.
Mar 2023: Young Jin Kwon joined our lab. Welcome!
Feb 2023: Our paper, "MicroDeblur: Image Motion Deblurring on Microcontroller-based Vision Systems", is accepted to IPSN 2023.
Oct 2022: Jaeho Kim joined our lab. Welcome!
Aug 2022: Jae Wook Lee and Kwangryeol Park joined our lab. Welcome!
Jun 2022: Jeongmin Cheon joined our lab. Welcome!
Mar 2022: Bosung Kim joined our lab. Welcome!
Feb 2022: Isu Jeong, Xiyana Figuera Michal, and Insung Jung joined our lab. Welcome!
Jan 2022: Changhyun Lim joined our lab. Welcome!
Dec 2021: Changhyeon Lee joined our lab. Welcome!
Dec 2021: Our paper, "Weight Separation for Memory-Efficient and Accurate Deep Multitask Learning", is accepted by PerCom 2022.
Aug 2021: Seulki Lee joined our lab. Welcome!
Page updated
Google Sites
Report abuse