Hi, I'm littlepeachs🍑

I work at the intersection of artificial intelligence, scientific discovery, and materials research. My current interests include molecular and polymer generation, AI-assisted materials design, scientific data representation, and agentic workflows for reproducible research.

I enjoy building research code that connects machine learning methods with real scientific problems: from molecular and crystal modeling to automated scientific-figure reconstruction and reproducible benchmark pipelines.

Research Interests

• AI for materials science and molecular discovery
• Molecular, polymer, and crystal generation
• Representation learning for scientific data
• Scientific agents for code generation, figure reproduction, and research automation
• Reproducible workflows for computational science

Selected Projects

• SurFF - A machine-learning force-field framework for predicting surface energy, surface exposure, and synthesizability of intermetallic crystals for catalyst discovery.
• TSFF - A machine-learning force-field method for efficient transition-state search and end-to-end automated catalyst screening.
• CatAI - A composable catalytic intelligence engine that fuses pretrained knowledge modules for catalyst property prediction, virtual screening, and closed-loop optimization.
• PolyRL - A reinforcement learning framework for generating polymer membranes optimized for CO2/N2 gas separation.
• SR4Aging - A symbolic regression framework for discovering interpretable relationships between microscopic features and macroscopic aging behavior in polymer materials.

Links

• Google Scholar

Contact

I am open to discussions on AI for science, materials generation, scientific agents, and reproducible computational research.

Collapse