I'm a third year Computer Science and Data Science double major at New York University. I like learning about brains both human and artificial.

A reinforcement-learning based approach to creating a chess bot. The name Chess Dreams comes from the use of a replay buffer to decorrelate state and simulate dreaming.

Lived experiences, like chess game states, are highly correlated with what comes directly before and after. Random selection from the buffer, like dreaming in nature, reduces the amount of redundant information absorbed. The model learns from separate experiences.

Scikit Learn, Pandas, NumPy, SQLAlchemy, RL

An exploration into AWS and a way for me to keep track of what I'm doing with my time on Earth.

Like the portfolio you're on right now, I built it using vanilla CSS. It feels good to keep things simple in a bloated tech landscape.

Feel free to browse as a guest or create an account and leave a little note in my guestbook!

AWS, HTML/CSS/JS, Cognito, DynamoDB, S3, Lambda

An NBA game predictor (also known as Hoops-Seer) that achieved an AUROC of 0.8 by training on a novel statistic: kills.

Generated by scraping through tens of thousands of NBA plays, "kills" track how often a team strung together three defensive stops in a row.

This project won First Place Overall at the NYU Data Science Expo.

Scikit Learn, Data Cleaning, Feature Engineering

A modeling system made for the Corner App designed to help users find restaurants and hangout spots based on four calculated "vibes": yum, glow, spark, and spice.

Calculated using dimensionality reduction into four axes, these labels numerically rank factors like restaurant quality, the best time of day to visit, and how adventurous a location's patrons would be.

This project won First Place Overall at the NYU DSC Datathon.

Scikit Learn, Pandas, Numpy, Google Colab