I’m currently a graduate student at UCAS (Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences). My current research areas include graph neural networks, AI for chemistry, and energy-based models.
My vision is to help accelerate the arrival of AGI and use technology to improve everyday life. I approach this from a scientific perspective—exploring AGI from first principles—while keeping a close eye on the latest advances and how to apply them in the real world.
I’m passionate about emerging technologies. Areas I focus on:
- AI
- Blockchain / Crypto
- Agents
Molecular Representation Learning Graph neural networks, self-orthogonalizing attractor neural networks, energy-based models
AI for Chemistry Predictive modeling of electronic structure (e.g., HOMO–LUMO gaps), electrolyte design, SEI interphase formation, reaction intermediates.
Graph Neural Network Pipelines for Molecular Properties Developing clean and reproducible pipelines combining RDKit, PyTorch Geometric, and Chemprop; conducting model ablation, hyperparameter tuning, and representation studies.
Agent Development I’m currently excited about agent projects and am experimenting with building a multi-agent system that collaborates across multiple projects.
Python, PyTorch, PyG, RDKit, Chemprop
LLM/Agent frameworks: LangGraph, Dify, custom tool-use pipelines
Data systems: scientific datasets (QM9, QM9-star, PCQM4Mv2, LIBE, MPcules)
Computational chemistry fundamentals: basic DFT workflows, electronic structure descriptors, reaction network analysis
My recent work includes literature reviews, structured reading notes, GNN derivations, and technical blog posts on molecular ML pipelines. I aim to make scientific computation more interpretable, reproducible, and accessible.
I aim to explore foundational deep learning architectures toward AGI. By mastering cutting-edge technologies, I hope to gain the freedom to explore technology-driven civilization.