Jaemin Park박재민
Assistant Professor, Department of Mathematics, Sookmyung Women's University
Seoul, Korea
About
I am an Assistant Professor in the Department of Mathematics at Sookmyung Women's University (since March 2026). My research sits at the intersection of pure mathematics and applied AI, organized along three axes. On the mathematical side, I work on homogeneous dynamics, Diophantine approximation, and topological data analysis (TDA). On the methodological side, I develop graph neural networks / graph transformers and generative models on manifolds (flow matching, diffusion, SE(3)-equivariant methods). I bring these tools to applications in electronic design automation (EDA) and semiconductor AI, simulation-driven molecular generation, and medical data science. Prior to joining Sookmyung, I was a Staff Engineer at Samsung Electronics AI Center.
한국어 소개
저는 숙명여자대학교 수학과 조교수입니다 (2026년 3월 임용). 순수수학과 응용 AI의 경계에서 연구하며, 세 축으로 구성됩니다. 수학적 기반 으로는 homogeneous dynamics, Diophantine approximation, 그리고 topological data analysis (TDA) 를 다루고, 방법론 으로는 graph neural networks / graph transformers 와 manifold 위의 generative models (flow matching, diffusion, SE(3)-equivariant methods) 를 개발합니다. 이 도구들을 응용 측에서 electronic design automation (EDA)·반도체 AI, 시뮬레이션 기반 분자 생성, medical data science 에 연결하고 있습니다. 숙명여대 부임 이전에는 삼성전자 AI Center 에서 반도체 설계·검증을 위한 AI 를 연구하는 Staff Engineer 로 재직했습니다.
News
- 2026-05-13 talk Invited seminar at Chung-Ang University, Dept. of Mathematics
- 2026-05-04 talk Invited seminar at Pusan National University — Geometry & Topology Seminar
- 2026-04-13 talk Colloquium at Sookmyung Women's University, Dept. of Statistics
- 2026-03-01 appointment Joined Sookmyung Women's University as Assistant Professor of Mathematics
- 2026-02-01 paper BADGE accepted at DATE 2026 (Verona, Italy)
Research Interests
Mathematical Foundations
- Homogeneous Dynamics — boundary measures on metric graphs and hyperbolic buildings, equivalence with Patterson–Sullivan and harmonic measures
- Diophantine Approximation — Hausdorff dimension of (weighted) singular vectors on fractal sets (Cantor-type)
- Topological Data Analysis — atlas flow, circular coordinates, persistence-based density-robust analysis
Geometric & Graph-based Learning
- Graph Neural Networks & Graph Transformers — positional/structural encodings, spectral and kernel-based methods (Green, Martin)
- Generative Models on Manifolds — flow matching, diffusion, SE(3)-equivariant methods, Riemannian geometry of conformer manifolds
- LLM Agents on Graph Tasks — loop-aware reinforcement learning (GRPO) for graph navigation
Applications
- Electronic Design Automation (EDA) — analytical placement, routing congestion, S-parameter prediction, timing optimization for semiconductor design
- AI for Molecular Generation — simulation-driven design with expensive reward signals (MD/FEP)
- Medical Data Science — network analysis of cardiac and neural signals, volume-entropy methods