xorq is a Python-based machine learning processing tool designed to simplify and accelerate your ML workflows. xorq brings together diverse data engines like DuckDB, DataFusion, and Snowflake under one unified interface.

Whether you’re a data scientist or ML engineer, xorq empowers you to focus on innovation rather than infrastructure.

​
Features

  • Expr (Ibis-Powered): Define transformations using a familiar, Pythonic API that is both expressive and backend-agnostic.
  • Cache: Avoid redundant computations by materializing intermediate results as Arrow RecordBatches, ensuring minimal re-computation.
  • Portable Python UDFs: Write and serve Python UDFs seamlessly – whether they’re used for aggregation, windowing, or transformation tasks.

Mission Statement:
“Make data processing ergonomic, performant and reproducible.”

​
Getting Started

The first step towards building declarative pipelines is to set-up your project.

Quickstart

Get a first taste of xorq

Core Concepts

The concepts behind xorq

​
Tutorials

Dive in to learn more about how to use xorq.

10-minutes tour of xorq

Learn key concepts in a brief tutorial

Pipelines as Expressions

Declarative, multi-engine expressions for ML

Build & Run

Build your expr pipeline into executable and introspectable artifacts

UD(X)F

Check out how to make it your own with extensible UDFs

Core Concepts
Next