What I work on
One question has followed me through several projects. Machine learning may predict a Hamiltonian quickly, but we still need to calculate the electronic structure, transport, or optical response we actually care about. At large scales, the bottleneck often moves from the physical model to the numerical methods and computing infrastructure around it.
That has led me to learned quantum-operator models such as DeePTB and SLEM; to distributed sparse-matrix software such as VBCSR; and to quantum-transport work including AD-NEGF and DeePTB-NEGF. They are different problems, but they connect.
Why I build tools
Coding agents can already read papers, modify code and run scientific jobs, but the scripts and logs they leave behind are often hard to make sense of. I started flower to preserve those steps as a workflow someone else can inspect and rerun. For me, a useful scientific tool should not only produce a result; it should also make clear how that result was obtained.
Background and papers
I received my bachelor's degree from Tiangong University and my master's degree from Tianjin University. I am now a Physics PhD candidate at McGill University, supervised by Prof. Hong Guo. My publications have appeared under the names Zhanghao Zhouyin, Yinzhanghao Zhou, and YinZhangHao Zhou.
The research notes explain what led to each project. The publication list keeps the formal papers, preprints, and citation details together.
Get in touch
If you work on electronic structure, quantum transport, scientific computing, or research agents, I am always interested in discussing concrete problems.