skfolio

Why we included this project

Quantitative analysts and systematic investors who need to turn return and covariance estimates into concrete portfolio weights will find skfolio practical. It bundles a wide range of allocation strategies, from classic mean-variance and risk-parity to hierarchical clustering and distributionally robust CVaR, behind a scikit-learn style API, so models can be cross-validated and stress-tested with the same workflow used for any other estimator. The built-in estimators for expected returns, covariance, and distributions, including copulas and shrinkage methods, let you assemble a full research pipeline without stitching together separate packages. Because it follows scikit-learn conventions, teams already comfortable with that ecosystem can adopt it quickly and integrate it into existing backtesting or research code. It is a good fit for anyone building reproducible asset-allocation and risk-management tooling in Python.