Concepts

Explore Xorq’s core concepts and learn how to build portable, multi-engine ML pipelines

Xorq is a unified platform for building portable ML pipelines that run across multiple compute engines. This section covers the fundamental concepts, architecture, and design principles that make Xorq different from traditional data processing tools.

Explore the documentation

Fundamentals

Learn about Xorq’s architecture, deferred execution, and expression format

Execution and backends

Understand how Xorq executes across engines, caches results, and manages connections

Reproducibility

Discover how Xorq ensures reproducible computations with hashing and versioning

Advanced capabilities

Explore UDFs, UDXFs, feature serving, lineage tracking, and batch processing