I'm a CS student building toward a career in software engineering, with a growing focus on applied AI. I'm currently expanding into Java and Spring Boot while deepening my understanding of how LLMs work and how to build practical applications around them.
📫 Reach me: LinkedIn
Keeper — AI Outfit Restyler Photograph an outfit that failed, mark the piece you love, and get three restyles that keep it — built from real in-stock clothing, rendered on your body, in colors matched to your complexion. Built for the YouCam API Skin AI & Apparel VTO Hackathon.
My role: product concept, workflow, data model, and UI design. Built with AI-assisted development (Claude).
🔗 View repo · Tech: Next.js · TypeScript · Prisma · Tailwind CSS · Vitest · YouCam Apparel VTO API · YouCam Facial Color Tones API
Contributing to GitLab as a community contributor.
- Completing my CS degree
- Expanding into Java and Spring Boot for backend development
- Building a portfolio that reflects real engineering skills
- Got into CS because I love the challenge of breaking down hard problems and building something real from the solution
- Self-taught pianist and guitarist
- Recently studied abroad in Madrid, Spain
Four projects from CodePath's AI 110 course, exploring debugging, scheduling, music recommendation, and LLM-augmented systems.
🕵️ Game Glitch Investigator — Module 1
A debugging exercise built on a broken Streamlit number-guessing game. Tracked down and fixed six bugs, including a resetting secret number, reversed hints, a broken score system, and a stale debug panel, then refactored the logic into a testable module with a full pytest suite.
Tech: Python · Streamlit · pytest · Debugging
🐾 PawPal+ — Module 2
A Streamlit pet care scheduler that builds a conflict-aware, priority-sorted daily plan from a task list. Features include time-budget enforcement, overlap detection across pets, daily/weekly task recurrence, and a validation layer that surfaces overbooking and assignment warnings.
Tech: Python · Streamlit · dataclasses · OOP
🎵 SoundMatch 1.0 — Music Recommender · Module 3
A content-based music recommender that scores songs against a user taste profile using weighted features: genre (40%), mood (30%), energy closeness (15%), danceability (10%), and acousticness (5%). Fully transparent, every recommendation can be explained by the scoring formula. Includes a model card analyzing bias toward exact genre/mood string matches.
Tech: Python · Content-based filtering · pytest · AI Fairness
🚀 SoundMatch 2.0 — AI-Augmented Recommender · Final Project
Extends SoundMatch 1.0 by placing two AI recommendation approaches side by side: a naive LLM call (Gemini 2.5 Flash) and a retrieval-augmented (RAG) call where the rule-based scorer acts as the retriever. Makes trade-offs concrete, including how RAG handles edge cases like missing genres that the naive LLM silently fumbles.
Tech: Python · Gemini 2.5 Flash · RAG · LLM Evaluation · pytest