On-going Projects
Seoul Business Agency (SBA) · Fintech Commercialization Program
ELS Hedging Decision System via Neural Optimal Control
Formulates the hedge-ratio decision for structured products as an optimal control problem, with development carried out by research teams at Sungkyunkwan University and Hanyang University.
TIPS R&D
Investor-Tailored Structured Derivatives via Time-Series Generation
Builds a recommendation and selection procedure for investor-tailored ELS with decision-focused learning (SPO+ surrogate loss), a capital-adjusted objective, and simulation-based screening of maturity exposure.
Research Grants
National Research Foundation of Korea · Research Grant
Robust Generative Representation Learning of Financial Data Using Vector Quantized Variational Autoencoder and Its Applications
National Research Foundation of Korea · Humanities & Social Sciences, Joint Research
Applications of Artificial Intelligence in Portfolio using Vector Quantization-Variational Autoencoder
Discrete latent factor models based on vector quantization; results published in Knowledge-Based Systems and at IJCAI-ECAI 2026.
National Research Foundation of Korea · Research Grant
Loan Eligibility Classification Model using Explainable Artificial Intelligence
National Research Foundation of Korea · BK21 Four
An Education and Research Team for Sharing and Cooperation based Smart Systems
Industry Projects
Korea Financial Telecommunications & Clearings Institute (KFTC)
Analysis of Financial Activities using AI Technologies
Defined units of analysis for financial activity data and grouped activity patterns by combining contrastive embeddings with unsupervised clustering.
FnDataLab
Machine Learning-Based Model Portfolios for Retirement Pension Management
Asset embeddings from self-supervised learning, turned into allocations for each investor risk propensity and evaluated by backtest.
Risk X tech Lab
Pricing of Exotic Options and Structured Products via Machine Learning
Deep hedging and neural-operator pricing models for ELS and other structured products, reducing the computational cost of conventional numerical methods.
FnDataLab
Machine Learning-driven Risk Rating Assessment of Global Investment Assets
Contrastive clustering of assets with similar risk characteristics, mapped to risk grades and documented in a validation report.
IOI Soft
Sales Analytics based on Graph Neural Network and Natural Language Processing
B2B transactions modeled as a graph, combining graph neural network representations with text embeddings of the related documents.
Hyperlounge
Business Analytics for Small and Medium-sized Enterprises
Organized the management data of small and medium-sized enterprises into an analyzable form and designed an analytics framework for decision-making.
Korea Forestry Promotion Institute (KOFPI) · Research Grant