Sean (Sihyun) Kwon
Data Scientist
New York, NY
adb0gv@r.postjobfree.com
github.com/seankwon1014
linkedin.com/in/seankwon1014
Columbia University, MS in Applied Analytics, GPA 3.92/4.0 Dec 2019, New York, NY
• Focus Areas: Machine Learning Theory with Statistics, Gradient Boosting, Deep Learning, Fraud Detection
• Core courses: Anomaly Detection, Capstone Project, SQL, Managing Data & Relational DB, Frameworks & Methods for ML Hanyang University, BS in Sociology, GPA 3.75/4.0 Feb 2008, South Korea
• Summa Cum Laude; Focused on social statistics to study social issues Education
Capstone Project - Data Science Team, Spruce Jun 2019 - Aug 2019, New York, NY
• Provided a foreclosure detection model by using Deep Learning with SMOTE; increased ‘recall’ score from approximately 20% of logistic regression to 72% of Deep Learning-FastAI Kaggle Competition
• Santander Customer Transaction Prediction: Top 11%, Binary Classification, Python, XGBoost & LightGBM
• Elo Merchant Category Recommendation: Top 11%, Regression, Python, XGBoost & LightGBM Projects
Summary
• Data Scientist with experience executing data-driven solution using deep learning, gradient boosting, and clustering, creating efficiency within marketing, product & service strategy, and anomaly detection
• Data Analyst with 5 years of experience analyzing smartphone-related online transaction data Skills
Machine Learning
Anomaly Detection: PyOD (DBScan, Meanshift, Autoencoder) Deep Learning: Pytorch / FastAI Gradient Boosting: XGBoost, LightGBM Regression / Clustering / Decesion Tree / Recommender System / NLP Data Visualization: Tableau/Matplotlib/Seaborn
Python: Numpy / Pandas / Sklearn / H2O
SQL / Hadoop
AWS: EC2 / Sage Maker R
Columbia University
Machine Learning Tutor Jul 2019 - Dec 2019, New York, NY
• Taught graduate students Python and SQL one-on-one at the School of Professional Studies Samsung Electronics Global HQ
Data Scientist Mar 2016 - Jun 2018, South Korea
• Marketing: Defined target customers based on ‘gain and lift chart’ by XGBoost; provided personalized content recommen- dation by collaborative filtering; improved download rates from 3% to 9% in the Galaxy Wallpaper Store
• Product Strategy: Created user clusters for Bixby, A.I assistant in Galaxy through K-means; provided insights for customer churn by delivering main used features and WAU by a cluster and period; decreased churn rate from approximately 45% to 30% in the next version
• UX: Provided UX pain points by detecting where users frequently lose their ways in the setting menu by analyzing UX log data on SQL; short-cut menu related to the analysis was added in the subsequent version Data Analyst Aug 2010 - Feb 2016, South Korea
• Analyzed sales trends by app categories and payment methods in the Galaxy Apps Store; contributed to adopting new pay- ment methods based on expected profits
Professional Experience