01 Home02 Research03 Projects04 Notes05 About06 Lab
Home/About

A physicist's curiosity,
a statistician's skepticism.

I am Liyu Cao, a PhD researcher in Statistics at the University of Hong Kong. I am interested in causal inference, statistical machine learning and the decisions we make from uncertain data.

I began in physics, where a model earns its place by explaining a mechanism and surviving a test. Statistics gave me a sharper language for uncertainty, and finance made the gap between a prediction and a decision impossible to ignore.

At HKU, I work on statistical research involving health-related data. In parallel, I build quantitative tools and study how financial models behave outside a polished benchmark.

Outside research:

PhotographyAerial photographyRunningSwimmingMixology

The Lab is where small statistical ideas become things you can play with. For research or collaboration, send me an email.

The path so far.

Newest first

Quantitative Research Analyst

Simplicity Capitalsince 2026

Quantitative research in industry, alongside the PhD.

Physics degree, Finance minor

Nankai University2021 – 2025

Bachelor’s training in physics with a minor in finance; National Scholarship recipient.

Optimization under uncertainty

UC BerkeleyResearch experience

Research experience in Ying Cui's group: a semismooth Newton approach to risk-aware optimization.

Deep learning price forecasts

Fudan UniversityJun – Sep 2023

Group research in Weiping Zhang's group: forecasts evaluated through an A-share backtest.

Active topological glass

Nankai UniversityNov 2022 – Apr 2023

Molecular dynamics research in Yao Li's group, presented at a national condensed-matter conference.

Training and tools.

Education
Bachelor’s in Physics with a minor in Finance, Nankai University; Berkeley Global Access exchange, UC Berkeley, 2023; doctoral study in Statistics, HKU
Recognition
National Scholarship, Nankai University
Teaching
Quantitative Strategies and Algorithmic Trading (STAT8020); Spatial Data Analysis (STAT6016)
Tools
Python and R for analysis; PyTorch and scientific computing for models; reproducible code and visual explanations