Research Scientist – Data/Machine Learning, NTT R&D | Japan                                                                                                                                Sep 2023–Aug 2024

• Developed and deployed an end-to-end pipeline using Python and TensorFlow, processing 2M+ sequential 5G throughput records with custom ETL for large-scale ingestion
• Engineered spatial, video and numerical features; conducted exploratory analysis (EDA) with encoding and embedding techniques to prepare model-ready datasets
• Built and tuned LightGBM and LSTM models for sequential data prediction; improving signal prediction accuracy by 11% through hyperparameters tuning and optimization
• Partnered with network engineers to interpret ML results and integrate predictive insights into 5G system simulations and testing workflows

Research Assistant of Crack Detection using ML, Georgia Tech | Atlanta, GA                                                                                                      Aug 2022–Present 

• Designed a machine learning system to support data-informed highway maintenance, including vehicle counting and crack detection from real-time surveillance cameras
• Annotated and preprocessed 100+ crack images; performed EDA and microfeature extraction to enhance model
• Compared and fine-tuned multiple computer vision models (MLP, CNN variants), batch-trained models in PyTorch, improving crack detection accuracy from 85% to 89%
• Developed data augmentation and feature engineering strategies to improve model interpretability

Sustainability Data Intern, Office of Campus Sustainability | Atlanta, GA                                                                                                        May 2022–Aug 2022

• Integrated multiple data source to build databases from scratch to deliver internal dashboard for compliance and public transparency from 28+ departments to support performance benchmarking
• Edited executive sustainability report and restructured internal pipelines to improve data reliability to stakeholders

Customer Churn Behavior Analysis Prediction                                                                                                                                                                    Aug 2024–Dec 2024

• Evaluated models (Linear Regression, K-Means, PCA) to identify high-risk churn segments and key behavioral drivers
• Enabled proactive customer retention through interpretable models (84% precision), supporting the revenue uplift

Full-Stack ERP System for Café Operations                                                                                                                                                                          Jan 2025–April 2025

• Designed and implemented a relational database and full-stack prototype ERP system using MySQL and Python Flask for café owners, supporting real-time inventory tracking, low-stock alerts, and customer purchase records
• Enabled supply chain optimization and proactive restocking by querying past procurement and automating supplier communication workflows