FeatureByte Inc., a startup working to simplify the task of developing artificial intelligence applications, today launched from stealth mode with $5.7 million in seed funding.

Glasswing Ventures and Tola Capital led the round.

FeatureByte’s co-founders, Chief Executive Officer Razi Raziuddin and chief product officer Xavier Conort, previously worked at AI startup DataRobot Inc. in leadership roles. They launched FeatureByte to ease an AI development task known as feature engineering. The task represents one of the most complicated aspects of machine learning projects.

In the enterprise, machine learning applications process a significant amount of data but often use only a portion of that data to carry out calculations. For example, an application tasked with determining the topic of business documents might only consult each document’s title instead of scanning the full text. The data points that an AI uses to make decisions are known as features.

An AI might require upwards of dozens of features to make decisions, which complicates software teams’ work. Moreover, providing machine learning applications with the features that they require often isn’t as simple as extracting the title of a business document. The task frequently involves complex data updates and calculations.

FeatureByte is working to ease AI development for enterprises. The Boston-based startup is developing a software platform designed to simplify the task of managing the features, or data, that an AI uses to make decisions. According to FeatureByte, its platform will speed up the preparation of data for machine learning applications and streamline various ongoing management tasks.

According to the startup, another task that its platform will simplify is monitoring. If erroneous records find their way into the data that an AI application processes, it may generate inaccurate decisions. To avoid such a situation, software teams use specialized tools to monitor their AI applications for errors.

FeatureByte reportedly plans to provide its software as a cloud service. Additionally, the startup will offer integrations with popular cloud data platforms such as Snowflake and Databricks.

“Our team has successfully launched AI deployments for hundreds of organizations worldwide,” Raziuddin said. “However, the one constant challenge enterprises face is feature engineering and management. Xavier and I formed FeatureByte to radically simplify the process for data scientists and application developers.”

FeatureByte intends to make its platform available to early users through an invite-only beta program. The startup will use its newly announced funding round to support product development, as well as accelerate go-to-market initiatives.

Photo: Unsplash