About

My research focuses on learning-augmented optimization, building on my background in temporal representation learning to connect sequential data and learned models with constrained decisions over time. I develop methods that combine machine learning with mathematical optimization to improve the efficiency and quality of decisions in large-scale constrained systems.

My current focus is large-scale, multi-period mixed-integer optimization, where discrete and continuous decisions are coupled across time and resources. I study neural solution prediction and structure-aware feasibility repair, with the goal of improving computational efficiency while maintaining feasibility and solution quality.

My master's research at Seoul National University focused on deep learning for time-series forecasting and anomaly detection, including temporal representations for long-range dependencies and multivariate structure. More broadly, I am interested in learning methods for optimization-based planning and sequential decision-making under complex constraints.