Open-source platform that teaches people how to use AI. 1.2M+ users, 42K+ discord community, 100+ contributors and 60+ modules.

Stabilizing Hostilities through Arbitration and Diplomatic Engagement

Built and deployed a software application using deep reinforcement learning for the effective pricing of The Graph protocol subgraph queries. Led the implementation of continuous policy gradient bandits (VPG, PPO) in Julia, optimizing adaptive query pricing.

Computer Science (Website Development)

The Farama Foundation is an ML research lab and non-profit that maintains nearly all of the largest open-source Reinforcement Learning libraries, including OpenAI's flagship project, Gym. Their tools are used by every major tech and AI research company (e.g., Google, Amazon, Microsoft, Meta, DeepMind, OpenAI) and have 50M+ installations.

Architected robust data pipelines to capture data from Arduino chips and RealSense cameras, efficiently integrating it with Microsoft Ψ. Designed and implemented a high-performing CNN+GRU model from scratch using Pytorch, substantially improving video classification accuracy on event camera data.Advisor: Andreas Andreou

Built infrastructure for a Diplomacy Discord Bot, enhancing player interactions with real-time authenticity predictions. Utilized Discord.js for seamless prediction displays. Designed a user-friendly website for data annotation initiatives and streamlined QANTA's data annotation workflow with Label Studio.Advisors: Jordan Boyd-Graber and Denis Peskov

Engineered robust unit tests for a project converting clustering algorithms from R to Python, enhancing code reliability. Developed scripts for algorithm performance visualization, bolstering data-driven insights. Advisors: Professor Joshua Vogelstein and Thomas Athey