Trading systems for people who read the backtest.
Open-source ML tools and commercial trading software, built by engineers who have shipped production data science for Walmart, Fiat-Chrysler, Deloitte, and Fidelity Information Services.
01 / Open Source · Active AlphaPy Pro
The modern Python 3.12+ ML framework for speculators. MetaLabeling, NLP sentiment features, MarketFlow pipeline, ensemble stacking across scikit-learn, XGBoost, LightGBM, CatBoost. Apache 2.0.
Python 3.12+ · Apache 2.0
02 / Open Source · Classic AlphaPy
The original Python ML framework, still up and downloaded 185,000+ times. A declarative YAML pipeline around scikit-learn, XGBoost, LightGBM, and Keras. Stable and supported.
YAML pipeline · 185K+ downloads
03 / Commercial TradeStation EasyLanguage
Our strategy code for TradeStation, ported from research notebooks and parameterized for live trading.
EasyLanguage · Live-ready
Hypothesis first
Strategies start as ideas, not backtests. We design the signal first, validate second.
Walk-forward only
No single-split train/test curves on this site. Every result is held forward in time.
Costs included
Published results include slippage, commissions, and the out-of-sample window.
Post-mortems shipped
If a strategy breaks, we publish the post-mortem. Survivorship bias is the enemy.
Audience
For systematic traders who want to read the code before they trust the curve.