Peer Reviewed Benchmarks - Redefining Feature Stores with Class Leading Performance ??

Our research paper, "The Hopsworks Feature Store for Machine Learning", is the first feature store to appear at a top-tier database or systems conference (SIGMOD 2024). In peer-reviewed benchmarks, Hopsworks was the class leading feature store, enabling the most challenging real-time AI systems, from personalized recommendations to financial trading.

In order to breakdown the results and findings from this research, we have created a series of articles aimed at describing, in lay terms, concepts and results from the study. The final part of this seven part series is out now!

Read The Final part: Reproducible Data for the AI Lakehouse

An explanation of how Hopsworks leverages its time-travel capabilities for feature groups to support reproducible creation of training datasets using metadata.

Read blog

Rest of the Story ??

Explore the implications of the paper in the following articles.

?? Part 1: Modularity & Composability for AI Systems - A presentation of a unified software architecture for batch, real-time, and LLM AI systems that is based on a shared storage layer.

?? Part 2: The Taxonomy for Data Transformations - An introduction to the taxonomy for data transformations in AI applications that is fundamental for any AI system.

?? Part 3: Snowflake Schema - An exploration of a snowflake schema data model for feature stores, and how it helps you include more features to make better predictions.?

?? Part 4: Lakehouse for AI - We cover the added value of a feature store over a data warehouse when managing offline data for AI.

?? Part 5: From Lakehouse to AI Lakehouse?- Read how Hopsworks generates temporal queries from Python code, and how a native query engine built on Arrow can massively outperform JDBC/ODBC APIs.

??? Part 6: RonDB - a Real-Time AI Database - Learn how Hopsworks (with RonDB) outperforms AWS Sagemaker and GCP Vertex in latency for real-time AI databases.