Personal Projects

Real-world projects, apps, and creative tools I’ve built to explore new technologies, test ideas, and sharpen my design and development process. Each one reflects my curiosity, experimentation, and focus on clean, purposeful interfaces.

CodeHub Website

FigmaResponsive DesignUI SystemWeb Accessibility

I led the UI/UX design and backend development for CodeHub, a student-run developer organization at the University of Georgia. The site serves as a central hub for forms, starter code, project showcases, and FAQs. I designed the full visual system and built the landing page frontend myself, while also contributing to backend architecture and helping coordinate integration work across the broader frontend. The focus was on creating a modular, accessible platform aligned with CodeHub’s mission to support beginner-friendly development and inclusive tech communities.

Case study not available

TrackifyJobs

Next.jsGoSupabaseStripe APITailwindCSSGemini API

TrackifyJobs is a job tracking software I built for tracking and refining my job applications. It combines NLP resume scoring, automated follow-ups, and a streamlined dashboard to reduce stress and increase offer rates. The goal was to make application management effortless and intelligent.

Case study not available

BoomBox - Mood-Based Music Recommendation App

Next.jsReactTailwindCSSMongoDBSpotify APIGemini APIPinataAuth0

BoomBox is a full-stack web app that recommends music based on the emotional tone of user-uploaded images. I led a team of two in building a system that uses Gemini for image sentiment analysis, the Spotify API for mood-aligned song selection, and Pinata for decentralized image storage. The app supports user authentication via Auth0 and stores user data in MongoDB. I designed the system architecture, handled API integration, and built the landing and core app frontend. Users can create and share playlists, with built-in group collaboration features that support co-curated music discovery.

Case study not available

E-Commerce Product Title Classification

PythonScikit-learnNLPTF-IDFSVCLogistic RegressionNaive Bayes

This project explores scalable machine learning methods for classifying e-commerce product titles into 248 categories using a dataset of over 1.4 million entries. I implemented a preprocessing pipeline with TF-IDF vectorization, stopword removal, normalization, and feature selection to the top 10,000 tokens. I evaluated multiple classifiers, including Linear SVC, Logistic Regression, and Multinomial Naive Bayes, with class-weighted adjustments to handle imbalance. The study showed that Linear SVC consistently performed best on large-scale product categorization tasks, demonstrating the viability of ML for text classification in e-commerce systems.

Case study not available