Investors flock to fund an AI cornerstone: Feature stores

Feature stores, with their dreary and opaque name, might not sound like the hottest subject.

They are an essential part of the artificial intelligence systems that enterprises and consumers use every day. They are attracting an increasing amount of attention and investment from venture firms, which see the market opportunity growing into the distant future.

Features are one of the components of an artificial intelligence system. Individual variables act like inputs in the system. In thinking about features, it is helpful to visualize a table where the data used by the system is organized into rows of examples and columns of attributes. Features are attributes used to describe each example, for example, a sender's contact information or words in the email body.

Working with features is an ad hoc process. At the enterprise scale, where data science teams are responsible for maintaining thousands of systems, a place to manage and track features becomes a necessity.

The feature store is a centralized repository for organizing, storing and serving the features that artificial intelligence systems rely on. feature stores are a unified place to build and share features across different teams in an organization.

Michael Del Balso, the CEO of Tecton.ai, said in an email that feature stores are at the intersection of data and machine learning. They are the interface between data and models.

Data engineers can see statistics on features, where they have been used and the impact they have had on models in feature stores. Users can aggregate, filter and join features without having to code. The feature value is the number of orders over the past 30 minutes.

Del Balso explained that advanced feature stores automate production to collect data from batches and real-time sources, transform the data in real time, and store it in the offline and online store. They often include built-in monitoring capabilities to keep an eye on the service levels.

The image is from Tecton.ai.

The feature stores promise to improve collaboration between teams. Tech giants and startup like Tecton are developing products to meet the need as the demand for them grows.