About Me
Professional Background
A seasoned AI Lead, Data Scientist, and PhD Economist with extensive experience developing and optimizing data-driven solutions in the B2C sector, including retail financial services and e-commerce. Specializes in applying advanced machine learning, causal inference (e.g., A/B testing, regression analysis, experimental design, RDD, DID), and structural economics to drive business strategy and product innovation. Proven ability to deliver significant business impact by translating complex data insights into actionable strategies for stakeholders, including C-level executives and heads of finance and operations.
Currently, as the AI Lead at VWS, I drive the product strategy for innovative solutions in the non-performing loan (NPL) sector. This role involves designing predictive models for NPL recovery, developing debtor segmentation for personalized repayment strategies, and prototyping new refinancing products. By integrating machine learning, behavioral analytics, and business modeling, I optimize loan recovery and build simulation tools to guide investment decisions in a highly regulated market.
Skills & Expertise
Programming & Tools
- Python (NumPy, pandas, Scikit-Learn)
- R and STATA for statistical analysis
- SQL for database querying
- Git for version control
Machine Learning
- Classification & Regression
- Deep Learning (TensorFlow, PyTorch)
- Clustering & Dimensionality Reduction
- Time Series Analysis
Data Analysis
- Exploratory Data Analysis
- Feature Engineering
- Large-Scale ETL
- Data Visualization (matplotlib, seaborn, plotly)
Causal Inference
- Experimental Design & A/B Testing
- Regression Discontinuity
- Difference in Differences
- Instrumental Variables
Education
Ph.D. in Economics
University of Southern California, 2021
Dissertation: "Essays on Causal Inference (Affirmative Action in Korea - Regression Discontinuity with Multiple Assignment Variables)"
M.A. in Economics
Korea University, 2010
Honors: Brain Korea 21 scholarship
B.S. in Mathematical Sciences
Korea Advanced Institute of Science and Technology (KAIST), 2008
Honors: Mathematical Science Department scholarship | National Science and Technology scholarship
Professional Experience
AI Lead / Data Scientist
Jul. 2025 - PresentVWS, Seoul, Korea
Leading the development of behavioral prediction and recommendation solutions to optimize NPL recovery
- At VWS, I lead data-driven product strategy for AI-powered financial solutions focused on non-performing loans (NPL). My role bridges machine learning, behavioral analytics, and business modeling to optimize loan recovery and design next-generation credit products.
Data Scientist
Oct. 2022 - Mar. 2024Datacrunch Global, Seoul, Korea (Remote)
Specialized in building data pipelines and demand forecasting modeling
- Causal-Inference Based Forecasting: Developed a hybrid model combining statistical intervention analysis with Deep Learning (LSTM). Causally separated organic demand from promotion-driven spikes to minimize inventory risk and quantify marketing ROI
- Scalable ETL Architecture: Designed AWS-based pipelines processing over 3M daily e-commerce transactions. Optimized PySpark jobs to reduce processing time by 70% and ensured data integrity across heterogeneous systems (WMS/ERP)
- Automated Decision Support System: Built a Data Warehouse integrating fragmented order/inventory data and deployed real-time ordering recommendation APIs to enable data-driven decision-making
Founder / Data Scientist
Oct. 2021 - Jul. 2025Decode Data Inc., Los Angeles, USA
- Startup Leadership: Led the entire product lifecycle from ideation to development and team building, establishing an Agile organizational culture
- ALM Valuation Engine: Developed an Asset-Liability Management (ALM) engine implementing bootstrapping algorithms for yield curve derivation and risk-neutral approaches for fair value calculation
Assistant Professor
Jun. 2010 - May. 2013Korea Army Academy at Yeongcheon (KAAY), Yeongcheon, Korea
- Defense Econometrics: Led government research projects on defense R&D efficiency. Estimated optimal defense spending using panel data regression and published findings in academic journals
- Organizational Analysis: Analyzed the correlation between economic incentives and organizational effectiveness using large-scale survey data to improve personnel management
- Educational Excellence: Designed economics curriculum and implemented a quantitative academic performance evaluation system, achieving the top departmental ranking for two consecutive years
Economic Researcher
Mar. 2009 - Mar. 2010Korea University (Client: Hyundai Mobis), Seoul, Korea
- Regulatory Risk Mitigation: Contributed to saving approx. $120M in potential fines by econometrically modeling pricing mechanisms to prove compliance with fair trade regulations
- Causal Analysis for Antitrust: Utilized Difference-in-Differences (DID) to analyze the causal effect of corporate actions on market prices, providing logical grounds for antitrust defense
Publications
Control Function Approach for Partly Ordered Endogenous Treatments: Military Rank Premium in Wage
Oxford Bulletin of Economics and Statistics, June 2017
Developed a novel statistical methodology to analyze the effects of military ranks on post-service wages, addressing endogeneity in rank assignment through an innovative control function approach.
View PublicationA Study on Scale of Defense Expenditure for Security Menace: A Panel Regression Analysis Approach
Journal of Korea Army Academy at Yeong-cheon, 2013
Led a government research project to design robust economic models for estimating optimal national defense R&D expenditure and efficient management, contributing to a larger project with a total value exceeding $40,000.
Affirmative Action in Korea - Regression Discontinuity with Multiple Assignment Variables
Dissertation Research, 2021
Developed an identification method for fuzzy regression discontinuity design with multiple assignment variables to analyze affirmative action effects in Korea. Found that while the overall policy showed no significant effect, company size-based implementation increased female employment rates by 5 percentage points.
Store Item Demand Forecasting Project
Technical Report, 2021
Implemented a Recurrent Neural Network with Long Short-Term Memory (LSTM) using Keras/TensorFlow to predict 3-month item sales across different stores, supporting business planning and cash flow management. The LSTM model reduced error rates to 86% of traditional ARIMA forecasting methods.
Customer Churn Prediction Project
Technical Report, 2020
Built a multi-classification model using XGBoost to identify customers likely to churn and determine the most influential features. The XGBoost model achieved an AUC of 93.3%, outperforming other algorithms including GBM (90.89%), Random Forest (87.76%), and Decision Trees (83%).
Online Retail Customer Segmentation Analysis
Technical Report, 2020
Segmented and cleaned business performance metrics including monthly revenue, activation rate, retention rate, and churn rate. Applied Lifetime Value (LTV) methods to improve multi-classification model accuracy from 76.5% to 84%.
A Study on the Estimation of Optimal Defense R&D Expenditure and Efficient Management
Korea Army Academy at Yeongcheon, 2012
Conducted comprehensive research on defense R&D expenditure optimization and management efficiency, providing evidence-based recommendations for resource allocation in national defense.
The Study on Scale of Defense Expenditure
Korea Army Academy at Yeongcheon, 2011
Analyzed defense expenditure patterns and requirements, developing models to determine appropriate funding levels based on security threats and national priorities.
Economic Effects of Alleged Anti-Competitive Behavior of Top Automobile Parts Company
Research Report for Hyundai MOBIS, 2010
Designed causal inference models to investigate economic effects of alleged anti-competitive behaviors on retail agencies, mediating companies, repair shops, and consumers. The research provided economic evidence that helped reduce the imposed fine from $150 million to $30 million.