Research & Industry Projects
Raon-VisionEncoder
Krafton AI
A Fully Open SigLIP2-class Vision Encoder
- Developed a vision encoder with comparable performance to SigLIP2-NaFlex using only open data
- Built VLM training pipeline integrating the vision encoder with a language model for downstream VQA evaluation and training optimization
VideoAlchemist 2.0
Snap Inc.
Multi-Subject Personalized Video Generation
- Developed a personalized video generation model supporting multiple subjects with fine-grained temporal control
- Built cross-reference dataset pipeline and multi-subject adapter architecture for personalized video generation
- Contributed to foundation of dataset generation pipeline and adapter design of AlcheMinT
3D-Aware VLM Finetuning
Samsung Research
- Electronic Device and Method for Operating Thereof and Storage Medium
- Paper: 3D-Aware Vision-Language Models Fine-Tuning with Geometric Distillation
Korean Patent: 10-2025-0109574
Personal Projects
CT Denoising CycleGAN
Low-Dose CT Image Denoising via Unpaired Image-to-Image Translation
- CycleGAN to denoise quarter-dose CT images into full-dose quality without paired training data
- Compared CycleGAN losses vs. supervised losses for CT image quality enhancement
Keywords: CycleGAN, Medical Image Denoising, Unpaired Image Translation
SHrack
Mobile Web Service for Real-time Exercise Count Tracking
- Web-based real-time video streaming service for accurate exercise count detection
Keywords: MobileNet, Contextual Prediction Module
KUVIS
Web-based Visual Programming Software with Deep Learning in Medicine
- Developed a prototype of VIENCE Pathoview
- Advised by Prof. Wonki Jeong · Jul. 2022 – Aug. 2022
Keywords: Visual Programming, Parallel Computing for Deep Learning