Starting lineup · No. 02 · Libero
Software engineer at Bank of America, Emory CS '26. I build tournament software, ML experiments, and tools for whatever I'm into. Lately that's a lot of volleyball.
Coach's board
Three touches per side. Dig, set, then spike for a clean kill. On the set, call your hitter on the court. Arrow keys work too.
Dig
Box score from every set so far. Emory University, B.S. Computer Science, 2026.
Software Engineer · Jersey City, NJ
Full time on a Java market-risk application, shipping monthly Value-at-Risk releases, maintaining a Python API for risk managers, and removing outdated Ehcache to unblock a Java upgrade. As an intern, built a VaR backtesting dashboard, prototyped a 7B+ record Iceberg migration on Spark, and shipped a Python charting library.
Software Engineer Intern · New York, NY
Worked on an AI course-support platform used by 20,000+ students and 20 professors. Shipped study-app features from Figma mockups and led WCAG 2.1 AA compliance.
Developer · Atlanta, GA
Case-interview prep platform with 10K+ users. Built the interview UI and voice-to-text transcription, and rewrote the GPT prompt templates that drive longer mock interviews.
Set
Set the ball, then call a hitter on the court. Flick sideways to run through them.
Go Outside · 2026
Full-stack tournament platform for collegiate club volleyball, covering registration, pools and brackets, court scheduling, and real-time live scoring. I built the tournament engine from scratch, plus a versioned REST API of about 90 routes shared by the web app and an Expo mobile client.
1 Middle quick · 2026
Research package and CLI that measures how much an LLM's refusals reveal about the hidden rules in its system prompt. Using only benign probes, every category leaked paraphrased policy content, and a second model that never saw the prompt reconstructed the intent of all three planted rules.
Red Opposite · 2024
Productivity web app built around the Pomodoro technique: timed focus sessions, shared breaks with friends, and a browser extension that blocks distracting sites. Built as a team project with a React front end, Auth0 sign-in, and a Flask and PostgreSQL back end on AWS.
Pipe Back row · 2024
Co-authored paper predicting NBA player point totals from 1997–2023 box-score data and using the predictions to estimate the expected value of player prop bets. Our tuned MLP regressor reached an R² of 0.651 and was benchmarked against random forest, ridge regression, and k-NN baselines.
3 3 ball · 2022
Research project under CMU professor Vipul Goyal forecasting next-day closing prices for Apple, Microsoft, Tesla, and Meta. I trained an MLP regressor on 100-day windows of daily adjusted prices from the Alpha Vantage API and evaluated it on a chronological 70/30 train/test split.