Computer Science & Mathematics at the University of Wisconsin–Madison

I build trading-system prototypes, machine learning systems, and quantitative research tools with an emphasis on reproducible evaluation, clear system boundaries, and measured results.

An independently developed C++20 price–time priority matching core with a flat price ladder, bitmap lookup, pooled FIFO queues, and an open-addressing order-ID map. Differential tests compare 1.6 million messages with a separate map reference. A single-thread Apple M4 synthetic replay reports 91.3M messages/s versus 16.1M for the reference; parsing and network I/O are excluded, so this is not production exchange throughput.

Reviewed version · Passing CI

C++20 · Cache-aware Data Structures · Differential Testing · ASan / UBSan

An independently developed, reproducible ten-factor study using the CSI 300 point-in-time universe. In-sample selection (2018–2021) is separated from evaluation (2022–2025 H1), with next-open execution, size neutralization, drift-adjusted turnover, and a survivorship-bias experiment. The composite records OOS RankIC 0.0324 and annualized ICIR 1.44. Locked dependencies, data checksums, and a frozen 100-stock regression sample support reproducibility. This is historical A-share research, not a live trading or U.S.-market performance claim.

Reviewed research report · Passing CI

Python · pandas / NumPy · Point-in-time Universe · Regression Testing

Inference engineering extensions for MiniMind, including static KV caching, dynamic request batching, and observable FastAPI serving. Reproducible CPU microbenchmarks compare cache and batching configurations while documenting their experimental limits.

PyTorch · FastAPI · Static KV Cache · Dynamic Batching · Prometheus

A leakage-aware modeling benchmark, SHAP evidence bundle, and validated reporting API. A fixed-parameter evaluation records CatBoost OOF RMSE 4.870 against a training-fold mean baseline of 6.890 across 430 eligible companies.

CatBoost · SHAP · scikit-learn · FastAPI

2025 Zhixiang Cup · Grand Prize
_{The award applies to the original competition case; the repository documents the subsequent engineering release.}

Machine learning systems · Applied mathematics · Interpretable modeling · Quantitative research

Personal website → · View all repositories →