Reinforcement learning
CritPT-RL
An end-to-end GRPO post-training and evaluation pipeline for scientific coding tasks, with synthetic data, reward functions, and executable verifiers.
Projects
Projects
Research tools, agent systems, and personal projects.
Research tools
README-AI
An AI paper reader for translating, discussing, and annotating research papers, with contextual figure explanations and related-work search.
Multi-agent systems
Texas-Poker-Agents
A locally runnable multiplayer LLM poker system for exploring agent behavior in imperfect-information games.
AI applications
Social Copilot
An iOS communication assistant that combines a message, relationship context, and writing preferences to suggest three replies for the user to review and copy.
Programming languages
Mini-Lisp
A Lisp interpreter in C++ with lexical scope, closures, macros, an interactive REPL, and script execution.
Interactive systems
OpenGame
A seven-day campus-life RPG about time, promises, and relationships, with a complete story playable locally without an API key.