jialong@columbia:~/site$cat ./lab/quant-dojo.md
> Lab · ink-dojo/quant-dojo

quant-dojo

repo:
ink-dojo/quant-dojo
lang:
Python · Streamlit
year:
2026 · active

quant-dojo is a systematic research environment for China's A-share market. It behaves more like a strict research room than a trading bot: ideas enter as factors, survive a sequence of tests, and only then earn the right to approach paper trading.

Why it exists

Quant ideas are easy to overfit and strangely hard for their authors to retire. I wanted the research to leave a legible trail: what was tested, what survived, what failed, and whether a signal still stood once it left the sample that flattered it.

What it does

  • Factor research tracks for A-share equities
  • Backtesting with walk-forward validation
  • Explicit research gates before an idea reaches paper trading
  • Risk controls and execution checks
  • Dashboards for monitoring research and live candidates

Stack

Python, pandas, Streamlit dashboards, broker and execution adapters, and a research log designed for reproducibility.

Repository: github.com/ink-dojo/quant-dojo.