Crunching data to catch fraud

About this talk

Veriff’s fraud team detects and blocks fraud by looking for patterns in the millions of incoming verifications.To be able to make good decisions in a fraction of a second, we need fast access to lots of data. Our automatic solutions can recognize known patterns of fraud, but we also need to manually add new rules on the fly when we recognise new attacks. In this talk I’ll show how we built a solution for quickly accessing that data at scale and running analytical queries on it. The result powers both our real-time decision making, as well as the tools for manually finding the patterns and blocking them.