Coupang is the first marketplace and largest online retailer in South Korea that generated $21 billion in revenue as of 2022. As a data scientist, I am actively involved in the growth of Rocket Growth, Coupang’s fulfillment service similar to FBA (Fulfillment by Amazon).
Team building - Established and led a data science/analytics team dedicated to Rocket Growth.
Pricing policy analysis - Employed price elasticity modeling and A/B testing to perform a comprehensive analysis of the pricing strategy.
Fee structure optimization - Conducted simulations and redesigned the seller fee structure for optimal performance.
Seller retention and growth modeling - Developed and implemented machine learning models to enhance seller retention and drive growth.
KPI development and dashboard design - Played a crucial role in developing KPIs and designing dashboards across the entire e-commerce business domain, covering areas such as seller acquisition, logistics, pricing, and sales.
Algorithmic trading - Designed and developed deep learning-based models for trading in commodity futures and cryptocurrency markets.
Portfolio management - Implemented a portfolio rebalancing algorithm tailored for multi-asset algorithmic trading, enhancing portfolio performance and mitigating risks.
Trade monitoring system - Constructed a robust real-time monitoring system specifically tailored for algorithmic trading, ensuring efficient and timely decision-making in a dynamic market environment.
Keywords:Algorithm trading, Deep learning, Trading system, Data visualization, Real-time analytics
Zalando is a fashion e-commerce company that generated €10 billion in revenue and had 46 million active customers in Europe as of 2020. I served as a data scientist in a competitive pricing team, focusing on pricing algorithm development and the analysis of sales data and business impacts.
Price optimization - Optimized revenue and profitability by assessing the effectiveness of competitive pricing through precise estimation of ROI, conducting monthly A/B tests for quantitative evaluation, resulting in a significant annual profit increase of 4 million euros.
Demand forecast & competitive pricing - Conducted a comprehensive analysis of competitors’ pricing strategy on sales performance utilizing cross-price elasticity of demand analysis, and developed a deep-learning model for sales forecasting to improve forecast bias for competitive products.
Product matching - Developed an ML pipeline integrating multi-modal embeddings and similarity calculation to match products from competitors’ websites with Zalando products, incorporating pre-trained deep learning models for enhanced embeddings.
Keywords:Revenue management, A/B testing, Data analysis, Machine learning
Lunit is a medical software company providing AI solutions that help discover cancer and predict cancer treatment outcomes.
AI solution for medical images - Developed CNN models for chest x-ray image analysis, surpassing the performance of experienced radiologists and improving radiologists’ performance as a second reader.
Keywords:AI medical imaging, Deep learning, Convolutional Neural Network
Pricing engine for financial derivatives - As a quantitative analyst in the FICC trading team, developed pricing models using Monte Carlo simulation, while also overseeing financial risk management and conducting scenario analyses.
Keywords:Financial engineering, Monte Carlo simulation
Tools:C++, SQL, Excel VBA
Ph.D. research topics / KAIST
High-frequency market analysis with deep learning - Developed and optimized a deep learning model for high-frequency market prediction, leveraging gradient analysis to visualize and analyze market feature influences.
High-frequency market making algorithm - Created and validated high-frequency trading algorithms using deep reinforcement learning, employing a simulated environment based on historical market data to ensure performance and stability of the trained policy.
Quantitative trader behavior analysis - Implemented a quantitative framework utilizing inverse reinforcement learning to measure and analyze the behavioral characteristics of market participants, examining how their behavior evolves in response to market conditions.
3D Ultrasound Image Reconstruction - Developed a robust framework for 3D ultrasound image reconstruction with Markov random field modeling to enhance the quality and accuracy of the reconstructed models.
Keywords:High-frequency market microstructure, Deep learning, Reinforcement learning
