DataVisor, a startup whose founders spent years working on computer security at Microsoft Research, is announcing today that it has raised $14.5 million in new funding.

Based on Apache Spark, DataVisor’s technology can analyze great quantities of activity logs — billions of events per hour — to identify fake or stolen accounts and prevent fraud, abuse, and spam.

The idea is to keep companies ahead of sophisticated attackers, who can test out and incubate attacks before launching them in order to maximize impact. “They’re not dumb out there,” DataVisor cofounder and chief executive Yinglian Xie told VentureBeat in an interview.

An increasing number of startups have come forward in recent years looking to cut down on unwanted usage and cyber attacks by collecting data at scale and employing machine learning. One such startup, Aorato, got acquired by Microsoft last year. More recently, Splunk bought Caspida.

But DataVisor is going after consumer-facing online services, not the Fortune 500, Xie said.

The startup will provide cloud software with a “threat intelligence console” and an application programming interface (API) to provide results of tests on its customers’ data, Xie said.

DataVisor started in December 2013 and is based in Mountain View, California. The startup employs 15 people. Customers include Momo and Yelp.

GSR and New Enterprise Associates both led the new round, which follows a seed round of an undisclosed size.