AiM Future, Inc.
AI Research Engineer · Advanced Research Team Seoul, Korea
Edge-AI & generative deep learning for vision
Hi! I'm Sean 👋 — I am currently working as an AI Research Engineer at AiM Future. My research centers on efficient & generative deep learning — diffusion & flow-matching models, quantization & knowledge distillation, and VLMs, VLAs, and LLMs — while optimizing AI models for on-device NPU deployment, especially computer-vision models.
Previously, I was an Applied Artificial Intelligence graduate student @ Sungkyunkwan University, researching in MAIN Lab, advised by Prof. Jaekwang Kim. I also was an SKT AI Fellowship member (4th generation) with the SKT Supernova team in 2022. I received my B.S. degree in Molecular Biotechnology @ Ghent University Global Campus and wrote my bachelor thesis under the supervision of Dr. Utku Ozbulak and Prof. Wesley De Neve.
Feel free to reach out if you'd like to chat or collaborate!
Industry Experience
SK Telecom
SK Telecom AI Fellowship — Supernova Team Pangyo, Korea
Deep learning for colorizing historical monochrome images (Team SKeTchbook)
Education
Sungkyunkwan University (SKKU)
Seoul, Korea
M.S. in Applied Artificial Intelligence
MAIN Lab · Advisor: Prof. Jaekwang Kim
Ghent University Global Campus
Incheon, Korea & Gent, Belgium
B.S. in Molecular Biotechnology
Thesis advisors: Dr. Utku Ozbulak & Prof. Wesley De Neve
Brent International School Manila
Manila, Philippines
High School · IB Diploma
Research Experience
MAIN Lab @ Sungkyunkwan University
Graduate Research Assistant Seoul, Korea
Advisor: Prof. Jaekwang Kim
Ghent University Global Campus
Undergraduate Research Incheon, Korea
Advisors: Dr. Utku Ozbulak & Prof. Wesley De Neve
News
- Jul 2026“Flow Straight to Reality: Perceptually Consistent Flow Matching for Efficient Image Restoration” accepted at ECCV 2026.
- Oct 2025First-author paper accepted as an Oral at the ICCV 2025 AIGENS workshop (Multi-Scale Contrastive-Adversarial Distillation for Super-Resolution).
- Mar 2025🏆 Our work won 3rd place in the 6th Thermal Image Super-Resolution Challenge (TISRC) — Track 1, at the 21st IEEE PBVS Workshop @ CVPR 2025. The code is open-sourced here and the paper is available here.
- Mar 2025“Debiasing Classifiers by Amplifying Bias with Latent Diffusion and Large Language Models” accepted at ACM SAC 2025 — presented in Sicily, Italy.
- Mar 2024Joined AiM Future as an AI Research Engineer.
- Mar 2024DiffInject accepted at the CVPR 2024 SynData4CV workshop. arXiv
- Mar 2024“Retrieval-Based Disease Prediction for Myocardial Injury after Noncardiac Surgery” accepted at the AAAI 2024 Spring Symposium on Clinical Foundation Models.
- Nov 2023A patent on hierarchical contrastive learning for sequential recommendation was filed in Korea.
- Mar 2023🏆 Best Paper Award (AI & Agents) at ACM SAC 2023 for MGCMA.
- Nov 2022🏆 1st Place & Minister's Award, 4th AI Grand Challenge (Phase 3, Track 1).
- Nov 2022🏆 Grand Prize, 4th SKT AI Fellowship (Supernova team).
- Mar 2022Began my M.S. in Applied AI at SKKU MAIN Lab.
Publications
* denotes equal contribution. See Google Scholar for the full, up-to-date list.
Published
Flow Straight to Reality: Perceptually Consistent Flow Matching for Efficient Image Restoration
European Conference on Computer Vision (ECCV), 2026
Multi-Scale Contrastive-Adversarial Distillation for Super-Resolution
ICCV 2025 Workshop (AIGENS), 2025 · Oral Presentation · Paper · Code
Debiasing Classifiers by Amplifying Bias with Latent Diffusion and Large Language Models
ACM Symposium on Applied Computing (SAC), 2025
DiffInject: Revisiting Debias via Synthetic Data Generation using Diffusion-based Style Injection
SynData4CV Workshop, CVPR 2024 · arXiv
AmpliBias: Mitigating Dataset Bias through Bias Amplification in Few-Shot Learning for Generative Models
ACM International Conference on Information and Knowledge Management (CIKM), 2023
A-ColViT: Real-time Interactive Colorization by Adaptive Vision Transformer
Practical Deep Learning in the Wild Workshop, AAAI 2023
Deepfake Detection for Facial Images with Facemasks
Workshop on Security Implications of Deepfakes & Cheapfakes, 2022 · arXiv
A²: Adaptive Augmentation for Effectively Mitigating Dataset Bias
Asian Conference on Computer Vision (ACCV), 2022
More publications
PacECG-Net: A Multi-modal Approach Integrating LLMs and ECG for LVSD Classification in Pacemaker Patients
ACM Symposium on Applied Computing (SAC), 2025
Hierarchical Contrastive Learning with Multiple Augmentations for Sequential Recommendation
ACM Symposium on Applied Computing (SAC), 2025
LMCG-Net: Integrating LLMs and ECG for LVEF-Based Classification for Pacing Patients
IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 2024
Retrieval-based Disease Prediction for Myocardial Injury after Noncardiac Surgery
AAAI Spring Symposium on Clinical Foundation Models, 2024
MGCMA: Multi-scale Generator with Channel-wise Mask Attention for Synthetic Contrast-enhanced Chest CT
ACM Symposium on Applied Computing (SAC), 2023
The Tenth NTIRE 2025 Efficient Super-Resolution Challenge Report
CVPR Workshops, 2025
Academic Service
Conference Reviewer
- BMVC 2025
- KDD 2024
- CIKM 2024
Patents
순차적 추천을 위한 계층적 대조 학습 방법 및 계층적 대조 학습 장치 (Hierarchical Contrastive Learning Method & Apparatus for Sequential Recommendation)
Patent filed in Korea · Nov 2023
Method and Apparatus for Generating Contrast-Enhanced Image from Non-Contrast Image using a Neural Network Model
U.S. Patent Application 18/603,467 · 2024
Community & Open Research
가짜연구소 (Pseudo Lab)
Text-to-Image Generation — Paper/Code Contributor & Learner Seoul, Korea
Reviewed papers & code on image generation and generative AI, and ran experiments fine-tuning diffusion-based models. Contributed to the community's GitHub.
모두의연구소 (Modulabs) — Robotics Lab
Researcher — Hierarchical VLA-RL for Humanoid Control Seoul, Korea
Hierarchical VLA–RL architecture for whole-body humanoid manipulation & locomotion (SmolVLA + RL; HumanoidBench, Unitree H1, MuJoCo).