Projects

Start with the implementations behind my reinforcement learning research, then explore earlier applied and learning projects.

Research implementations

ICLR, 2026. JAX / Python.

A Stackelberg formulation lets representations adapt faster than value estimates, stabilizing their interaction in deep Q-learning.

IJCAI, 2026. JAX / Python.

A pretrained world model communicates reusable context, helping a separate control agent learn faster and adapt to new tasks.

Collaborative research · Scientific Reports 2024

Collaborative work on detecting AI-generated and AI-revised graduate application documents, with a public codebase.

Text classification · Domain-specific evaluation

Independent learning project

A neural-network-based reinforcement learning agent that learns to play Snake. A compact implementation of the environment–agent learning loop.

Reinforcement learning · Neural Q-function

Data mining course project

A bank marketing classification project predicting whether a customer will subscribe to a term deposit, with code, report, and slides.

Classification · Model evaluation

Applied data analysis

Analyzes NYC collision records from 2016–2021, comparing injury and fatality patterns by time, borough, and contributing factors.

Exploratory analysis · Tableau

Algorithms & game theory study

An investigation of stable matching and its applications, presented through a written report and slides.

Stable matching · Game theory

Independent learning project

A recommendation systems learning project using Netflix competition data, with the implementation available on GitHub.

Recommendation systems

Independent study · In progress

An exploratory implementation study of text-to-image diffusion models, inspired by Imagen. This work is in progress.

Text-to-image diffusion

More code on GitHub.