• Developed Variational LSTM and Transformer to learn compact representations of high-dimensional, time-series data for Variational Inference• Leveraged parallel computing and multi-threading to reduce run time of probabilistic simulation written in c++ by 99%, decreasing it from 7000 seconds to 23 seconds• Built mathematical and Bayesian statistical models to explore dynamics of immune system

• Develop an interpretable machine learning model to analyze the impact of economic conditions on fraudulent activities in financial markets

Teaching Assistant for IEOR E4573 (Natural Language Processing), CS 4995 (Applied Deep Learning)

A MedTech company offering personalized medical devices for snoring and sleep apnea, backed by MashUp Angels and Softbank• Secured $600K investment to enhance company software development and data science capabilities by producing technical content for presentations at Tech Incubator Program for Startup Korea (TIPS)• Trained & deployed prototype CNN model using PyTorch/TorchServe to track snoring from audio signal, validating proof-of-concept (POC)• Automated e-commerce data pre-processing tasks using Apache Airflow, cutting processing time, operational costs, and human errors• Created an AWS Lambda/SMS API stack that sends thousands of texts a week, providing customers with real-time order updates and cutting down customer inquiries by 16%

Teaching Assistant for MAT100 (Calculus I), MAT102 (Calculus I), MAT103 (Calculus II), MAT104 (Calculus I) and MAT105 (Calculus II)