I was happy to participate in the final round of the SK hynix AI Hackathon 2026, where I was able to experience Forward Deployed Engineering (FDE), powered by Cofa Probe from KRAFTON. I was also honored to receive the AI Challenge Award (Top 20)!
It was a new and rewarding experience to build AI-powered solutions for real-world problems in the semiconductor industry. At the same time, I learned a lot from the people there, and the keynotes from SK hynix and OpenAI were truly insightful.
Our paper, InstantHand: Feed-Forward 3D Gaussian Splatting for Animatable Hand Avatar Reconstruction from a Single Image, has been accepted to NeurIPS 2026! [Paper][Code][Project Page]
This paper presents a method for building an animatable 3D Gaussian hand avatar from a single image in one forward pass, without per-subject optimization. It builds an avatar within seconds, renders in real time, and achieves quality comparable to per-subject optimized methods.
Our paper, MathLoopBench: Evaluating Agentic Mathematical Discovery Loops, has been accepted to the 6th Workshop on Mathematical Reasoning and AI (NeurIPS Workshop MATH-AI) 2026! [Paper]
This paper presents a benchmark for mathematical discovery loops, which must decide what is worth proving before proving it. Each loop grows a small Lean 4 library under a fixed budget and is scored by how many hidden target theorems it recovers.
I graduated from UNIST with a B.S. in Computer Science and Engineering and a minor in Mathematical Sciences! I was honored to receive the Summa Cum Laude distinction and the Deputy Prime Minister and Minister of Science and ICT Award.
I was fortunate enough to learn and experience so much with such wonderful professors, colleagues, and friends, and I am sincerely grateful for every moment 🥰.
Our paper, BIGS: Bimanual Category-agnostic Interaction Reconstruction from Monocular Videos via 3D Gaussian Splatting, has been accepted to CVPR 2025! [Paper][Code]
This paper presents a method for reconstructing bimanual object interactions from monocular videos. We propose a category-agnostic approach using 3D Gaussian Splatting to model both hands and objects without relying on prior object templates.
