Published
Writing Archive
Browse the current post archive and published long-form notes from the dedicated blog backend.
Open archiveBio
I am a first year Ph.D student in the Department of Computer Science and Engineering at Seoul National University, advised by Jaesik Park. I was also a research scientist intern at SONY Research, hosted by Takashi Shibuya and Masato Ishii. My research interests lie at the intersection of computer vision and machine learning, with a focus on generative models and multimodal understanding.
Full bio ▼I received my M.S. degree from Seoul National University, advised by Jaesik Park, and was honored with the Best Thesis Award. During my master's, I worked as a research scientist intern at Microsoft Research Asia with Chong Luo and Qi Dai.
I earned my B.S. in Electrical Engineering and Computer Science from Gwangju Institute of Science and Technology (GIST). During my bachelor's, I was fortunate to be advised by Jonghyun Choi at Seoul National University and Jeany Son at POSTECH. I was also a recipient of the Presidential Science Scholarship.
Latest News
Selected Papers
Talks
Reviewer Service
CVPR 2026, ICLR 2026, ECCV 2026, NeurIPS 2026, TPAMI 2026.
Academic servicePosts
Research notes, technical writeups, and longer-form posts.
Read postsProject
Demo services, course tools, and small deployed prototypes.
View projects
Upcoming research internship at Adobe Research, focusing on visual computing and generative modeling for controllable visual content creation.
Working on visual generative modeling and multimodal understanding, with an emphasis on controllable video generation and evaluation. The internship connects model behavior analysis with practical generation pipelines for user-facing visual media applications.
Investigated video generative models, subject-driven customization, and practical systems for open-domain customized video generation. This work led to an ICASSP 2026 paper and the arXiv 2504 preprint, Subject-driven Video Generation via Disentangled Identity and Motion.
Built an early research foundation in computer vision, visual recognition, video understanding, and generative modeling. Published multiple domestic conference and journal papers through KSC, KIMST, and the Journal of KIISE, along with international conference papers including ICCV 2025 and ICCV 2023.
I was grateful to collaborate closely with:
Master's thesis: Advancing Image Editing through Layer-wise Memory in Diffusion. The thesis studied controllable image editing in diffusion models through layer-wise memory mechanisms and was recognized with the Best Master's Thesis Award. Thesis link will be added later.
Took upper-division and theory-oriented coursework including EECS 126, convex optimization, CS 70, and discrete mathematics, along with related courses in probability, algorithms, and computer science foundations.
Graduated 2nd in the Department of Electrical Engineering and Computer Science and received Magna Cum Laude honors. Coursework and undergraduate research focused on machine learning, computer vision, and systems foundations.
* indicates equal contribution. † indicates corresponding authors.
Published notes stay public. The private writing studio now uses the shared admin session and the unified Mac backend gateway.
Published
Browse the current post archive and published long-form notes from the dedicated blog backend.
Open archiveAdmin
Sign in, draft posts, publish entries, and manage the local writing backend without touching other services.
Open private studioProject Access
Leaderboard
2026.03 ~ Present
Submission and ranking dashboard for the 2026S Introduction to Machine Learning assignment.
Open leaderboardPrivate Project
2026.04 ~ Present
Telegram-like web chat hosted on the Mac mini with a single shared admin account and one managed gateway entrypoint.
Open PocketChatPrivate Project
2026.04 ~ Present
Private group daily video check-ins with same-moment multi-view slots, invite-only sharing, and an end-of-day recap built for close circles.
Open Pocket LogPrivate Project
2026.04 ~ Present
Private writing studio backed by the Mac mini, now routed through the same managed gateway as the other private services.
Open Blog Studio