Lightning-fast data preprocessing and feature engineering for Python, built on Polars. 108 sklearn-style transformers for imputation, encoding, scaling, outlier clipping, discretization and feature selection, with Pipeline support and ONNX export for low-latency inference.
