Quantitative Risk · Analytics · Columbia MSBA '27

Modeling risk
under uncertainty

Retention Risk (Toyota) → Revenue Risk (American Airlines) → Portfolio Risk (Next)

Dual-degree senior at Michigan (Business + Data Analytics), incoming Columbia MSBA. One consistent thread: risk modeling — workforce retention risk at Toyota, revenue risk at American Airlines this summer, portfolio risk next.

View Projects Get in Touch
3
Internships
6+
Risk & ML Projects
2
Degrees
Yujin Kim
About

At Toyota, I built predictive risk models on 26K+ corrective-action records — identifying high-risk employee segments and quantifying retention risk with survival analysis and XGBoost. This summer at American Airlines, I'll apply scenario-based demand modeling under uncertainty to optimize pricing and revenue risk across competitive markets.

On my own, I engineered a portfolio risk simulation system inspired by BlackRock's Aladdin — computing VaR across Monte Carlo, parametric, and historical methods with Cholesky-decomposed correlated returns. Next: Columbia's MSBA, deepening stochastic processes, Bayesian modeling, and ML applied to financial risk.

Risk & Quant Methods
Monte Carlo Simulation VaR / CVaR Stress Testing Scenario Analysis Cholesky Decomposition Demand Forecasting
Programming
Python SQL R NumPy / SciPy pandas scikit-learn
ML & Analytics
XGBoost Random Forest Logistic Regression NLP Power BI Survival Analysis
Projects

Selected Work

I built a portfolio risk engine to understand how multi-asset portfolios behave under uncertainty. But I realized risk metrics alone don't help with decisions — so I extended it into an optimization system that suggests how to reallocate portfolios under constraints.

Phase 1 — Risk Modeling
Portfolio Risk Engine

End-to-end portfolio risk analytics system inspired by BlackRock's Aladdin. Computes VaR across Historical Simulation, Parametric (Gaussian), and Monte Carlo with Cholesky-decomposed correlated returns across 65+ assets. Includes stress testing against 6 historical crisis scenarios, correlation analysis, risk decomposition, and plain-language risk diagnostics.

3
VaR Methods
65+
Assets
6
Crisis Scenarios
Try Live Demo GitHub →
Interactive Demo Run Monte Carlo simulations,
view VaR outputs, stress-test
portfolios in real time
Phase 2 — Decision Optimization
Portfolio Decision Optimizer

After building the risk engine, I discovered that risk metrics alone were insufficient for decision-making — extending the system into an optimization framework that translates risk insights into actionable allocation strategies with risk tolerance constraints and return objectives.

4
Objectives
3
Constraints
10K
Simulated Portfolios
Try Live Demo GitHub →
Interactive Demo Efficient frontier visualization,
constraint-based optimization,
before/after comparison
02
Classification Risk
Bank Customer Churn & Risk Modeling

Classification models quantifying attrition risk across 10K+ accounts. Identified key risk drivers and modeled economic viability — projecting ROI even at 50% success rate.

10K+ accounts
Demo
03
Ensemble ML Stacking
Telemarketing Outcome Prediction

Stacked ensemble (LR, KNN, ANN, SVM, DT) on 41K+ records. Identified economic risk predictors and translated outputs into resource allocation strategy.

41K+ records
Demo
04
Optimization Pricing
CitiBike Dynamic Pricing Strategy

Analyzed rider demand, weather, and station capacity in NYC. Recommended dynamic pricing model to optimize utilization and stabilize revenue.

NYC network
Demo
Background

Three countries,
one analytical lens

Growing up between Korea and New Zealand taught me to translate across contexts — languages, cultures, ways of seeing. That same instinct drives how I approach data: finding signals that cross boundaries and building models different stakeholders can trust.

Seoul, Korea
Early years · Korean & English
Auckland, New Zealand
Auckland International College · IB Diploma · 2018–2021
Ann Arbor, Michigan
University of Michigan · BBA + BSI · 2022–2026
Dallas–Fort Worth, Texas
American Airlines · Revenue Management · 2026
New York City
Columbia University MSBA · 2026–2027
Education

Academic Foundation

Columbia University (Incoming)
M.S. in Business Analytics
Concentration: Algorithms & Financial Analytics
Expected Dec 2027 · New York, NY
Machine Learning Deep Learning Bayesian Modeling Stochastic Processes Experimental Design
University of Michigan, Ann Arbor
Dual Degree
BBA, Ross School of Business · BSI, School of Information (Data Analytics)
May 2026 · Ann Arbor, MI
BBA 3.92 · BSI 3.81
Predictive Analytics Corporate Strategy Data Manipulation Statistical Modeling
Contact
Let's connect

Open to conversations about quantitative risk, model risk, and portfolio analytics roles. Currently targeting Summer 2027 internships during Columbia MSBA.

[email protected] LinkedIn ↗ Resume ↓
© 2026 Yujin Kim · Ann Arbor → New York