Yihein Chai, a fifth-year medical student at University College London (UCL), maintains a personal website that reflects deep interests in neurosurgery, machine learning architecture research, and AGI. The portfolio serves as a central hub for Chai's work at the intersection of clinical medicine and artificial intelligence.

EpicGym and Reinforcement Learning

Among the highlighted projects is EpicGym, described as a local recreation of the Epic EHR interface. It is designed as a deterministic reinforcement-learning environment featuring synthetic charts, scored workflow tasks, and adapters for Gymnasium and computer-use. This tool aims to provide a structured simulation for developing and testing AI agents within medical data contexts.

Research Taste Lab

Chai also hosts the Research Taste Lab, a collection of 200 exercises designed to practice AI research judgment. These exercises include dated evidence, staged hints, and explanations tailored for both accessible and research-literate audiences. The initiative seeks to cultivate critical evaluation skills in the rapidly evolving field of AI research.

Writing and Interdisciplinary Analysis

The site features an extensive writing archive with posts dating back to 2018. Topics range from technical subjects like 'Generators, Iterables and Iterators' to philosophical and biological analyses such as 'Subclonal capitalism' and 'NMDA receptors: State management system in the brain'. Other notable entries include 'Continual Learning Through Introspection' and 'Similarity of System Design in Software Engineering and Medicine'.

Key Takeaways

  • The portfolio centers on Chai’s dual focus on neurosurgery and ML architecture research, including AGI interests.
  • EpicGym offers a deterministic RL environment for EHR interfaces, incorporating synthetic charts and scored tasks.
  • The Research Taste Lab provides 200 exercises with staged hints to improve AI research judgment.
  • The writing archive spans from 2018 to 2026, covering interdisciplinary topics linking biology, philosophy, and software engineering.

The Bottom Line

Chai’s portfolio distinguishes itself through a rigorous, interdisciplinary approach that treats medical concepts and software engineering principles as parallel systems. It offers valuable resources for developers interested in the specific niche of healthcare AI and reinforcement learning simulations.