Computer Science & Economics @ Harvard
Interested in game theory, quantitative finance, artificial intellignence, and automated insulin delivery systems
Red/Black — Counting and Solving a Hidden-Information Game
A three-paper series on Red/Black, a hidden-information bidding game played over a shared 52-card deck in which only color matters.
📄 Read the papers → · 🎮 Play the game →
Co-authored a closed-loop RL pipeline that fine-tunes Qwen3-8B against the Lean 4 compiler as an exact, unhackable reward — no learned reward model to game. GSM8K accuracy 79.0% → 86.5% (n=200, p=0.0046), with no measured regression across five alignment and sycophancy benchmarks.
Read more
I have worked on Loop, an automated insulin delivery system, both as an employee at The Tidepool Project, where I developed new features for the FDA-approved system Tidepool Loop, and as an open-source contributor to DIY Loop, which is used by more than 20,000 patients around the world. This project represents much of my work on GitHub and is central to who I am. Loop has not only improved my life by easing the burden of type 1 diabetes but has also shown me the extraordinary power of code to create meaningful, far-reaching impact.
Read more
These apps, published on the Apple App Store by Noble and Greenough School and Appier LLC, were programmed by me and are available free of charge. They include a comprehensive platform designed to support students and teachers, a daily journal and mood tracker that promotes happiness, and a social game where users test how well their friends know them. Collectively, the apps have surpassed 8,000 downloads, accumulated over 2 million sessions, and maintain hundreds of daily active users. Though they span diverse categories, each was created with the shared goal of making life a little brighter for others.
Read more
I value open-source projects as a way to give back to the community that has helped me throughout my coding journey. These apps are my original creations that have not made it to the App Store. Many of these projects are AI-focused, as it is a particular interest of mine within the field of computer science.
| 🏃 ActivityDetector | Classifies movement types from iPhone motion data with a neural net tuned across layer counts, widths, and temporal input strategies |
| 💰 NoahCoin | An ECDSA-signed cryptocurrency prototype and backend server, built from scratch in Python |
| 🏛 PikesvilleOpportunity | Econometrics paper on opportunity and place, written for Ec 50 |
| 🗺 HideAndCSeek50 | A geolocation hide-and-seek game inspired by Jet Lag: The Game |
| 🏫 NoblesClubsDirectory | A Vue web app live in production for student organizations |
| 📰 TheCrimsonApp | An iOS article reader for thecrimson.com |
| 🗳 TidemanVisualizer | Visualizing the Tideman ranked-pairs voting method |
Python · C · SQL · TypeScript · Java · Swift · Kotlin · TeX
CFR / CFR+ · equilibrium computation · Monte Carlo · statistical inference · distributed compute