Managed and enhanced ComptoxAI, an AI tool for toxicology data analysis, at Romano Lab. Key tasks included migrating to AWS EC2 with NGINX, automating processes using Bash, and developing Python and Neo4j features for data querying. Focused on modernizing the codebase and ensuring quality through standardized testing and issue management.

Led the launch of YodaTrip, an automated travel itinerary service app with over 200k downloads in South Korea. As the project manager, established the core functions and design concepts of the app using ProtoPie and Adobe XD. As the frontend developer, materialized the prototype into an app using TypeScript and Tailwind CSS. Contributed significantly in the design and development of the landing page.

Developed for the paper Adaptive Preference Measurement with Unstructured Data. This adaptive survey framework uses Bayesian optimization to measure consumer preferences from unstructured data like text and images. The program includes backend models and data processing scripts, focusing on real-time operations and automatic attribute analysis.

A REST API bookstore app using Django and Python, enabling self-publishing and sales of books. The project includes a custom user model with author pseudonyms, a comprehensive book model, and JWT authentication. Features include CRUD operations, endpoint security, and specific user restrictions.

A Kanban board that has all the basic functions, including task creation, editing, and filtering.

A very simple fullstack web application for file uploads, data storage, and file processing using AWS services. Built with React and managed using AWS Cloud Development Kit, the app features a workflow with AWS Lambda, S3, DynamoDB, EC2, and API Gateway.

A project utilizing Machine Learning and sentiment analysis techniques to analyze food reviews and recipes from Food.com. The project includes data cleaning, exploratory data analysis, model building, and visualization, all demonstrated in an IPython Notebook. It aims to provide insights into food preferences and recipe categorization.

Redundit is a Reddit-inspired forum community website, developed as part of the CIS-557 course. The project showcases full-stack development capabilities, focusing on React for frontend development and MongoDB for the backend. Key features include user authentication, profile management, group administration, post and comment system, and an interactive messaging window. The website also supports functionalities like group suggestions, hashtag filtering, and multimedia message.

Conducted an in-depth analysis of a five-year sales dataset for a major Consumer Electronics Company in the European market. Employed R for predictive modeling and data visualization, focusing on sales prediction, market basket analysis, and consumer segmentation. Uncovered unique purchasing trends and redefined customer segments. Also conducted apriori analysis for strategic product pairing. Contributed to securing 1 million USD in funding for a data consulting project with this Fortune 500 firm.