Efficient In-Memory Processing

Built with Rust, FeatureExpress offers a high-performance in-memory engine, enabling fast calculations and minimizing latency. Benefit from performance tricks like incremental updates of overlapping aggregates.

Powerful Time-based Queries

Design complex temporal queries and calculate features over predefined intervals, fixed dates, conditional dates, or custom observation times. Enjoy the clarity and robustness with a separation between past and future to avoid data leaks.

Flexible Domain-Specific Language (DSL)

Express your feature engineering logic using FeatureExpress's own DSL. Tailored for data scientists, the DSL provides a clear syntax to define a wide variety of queries, including time-based JOINS, aggregation functions, and filtering criteria.

Versatile Value Representation

FeatureExpress supports a broad array of value types including numbers, strings, dates, and complex structures like maps and vectors, allowing you to handle various data formats and create rich feature sets.

Convenient Indexing

The in-memory event store is equipped with multiple indices, including global and entity-specific indices, ensuring optimal retrieval and efficient management of event data.