eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Fraud prevention technology is commonly used in financial services, though there are wider use cases as there are a lot of different types of fraud.
In 2016, online IT career site Dice had a problem dealing with a large volume of fraudulent resumes and spam profiles. In fact, as much as 30 percent of account activities were fraudulent, and the company was challenged with finding a solution. Now in 2017, the volume of fraudulent spam profiles and account activity has been reduced to less than 5 percent, thanks to the use of machine learning adaptive fraud technology from security startup Simility.
Simility’s anti-fraud platform became generally available in May 2016 and was built by former Google abuse and ad fraud engineers.
Richard Maldonado, senior manager of product management at Dice, started at the company a year and a half ago and was tasked with rebuilding the candidate experience on the IT career site. He quickly realized a big problem: there was a high volume of fake candidates, which was causing disruption. The fraudulent activities included mass applications by fraudulent candidates for jobs and fake resumes flooding the Dice career database.
“One of the things we wanted to do was implement a solution where we would could identify fraudulent users and then take action to limit or restrict their impact on our system,” Maldonado told eWEEK.
Organizations are creating fake accounts on Dice that specifically target key search terms that employers might be looking for, he said. When a potential employer contacts the fake account, the organization that created the fake account will say that the person on the resume is not available but it has another great candidate.
“So there is some bait-and-switch happening,” Maldonado said. “There are also IT solution providers that are seeding our database with profiles, using them as lead generation to sell their services.”
To fix the problem of fraudulent accounts, Dice looked at several different technologies and ended up trying out Simility in September 2016.
“After six months with Simility, we were able to reduce the fraudulent activity to less than 5 percent,” Maldonado said.
Machine Learning
The Simility anti-fraud technology is able to help Dice’s operations, in-line and in real time, reduce and identify fraudulent activities.
“We have a number of check points where we capture event data and send it to Simility for analysis, and then we also have some control points where we request a decision from Simility on the status of a user,” Maldonado said.
The Simility platform is a machine-learning platform for helping to detect anomalous activities. Simility can understand what the activity of a typical normal user behavior should be and what indicators might indicate abnormal behavior. Simility’s platform uses multiple tools, including the open-source Elasticsearch project, to help find potentially fraudulent activities.
“When many people first think about fraud, they thing about credit card or banking fraud, but there are many other types of fraud that occur,” Rahul Pangam, CEO of Simility, told eWEEK. “There is an education element to our business that we focus on and use cases like Dice help us to push that message.”
