Banking: Real-time Fraud Detection

Financial services

Like all banks, one of Belgium’s biggest banks is constantly faced with new fraud threats whereby criminals are trying to steal from their unsuspecting clients. In addition, the European enforcement requires banks to provide instant payments to their customers.

Ask for a Demo

Challenge

New fraud threats

emerging constantly

Real-time analysis

needs to happen in <350ms

Solution

1. We built a real-time fraud detection app

3. And provides them with advanced capabilities without having to deal with coding

2. That is easy to use for the bank’s data scientists

4. So that they can swiftly apply the most current business requirements themselves

“Things work out best for those who make the best of how things work out.”

— John Wooden

Results

37
Milion+ Transactions per day in Belgium with...
49
of payment transactions
350
the average human eye blink lasts 100-400ms

Ask for a demo!

We’re your one-stop cloud-native partner

Big Data engineering. Data science. Data Analytics. Site Reliability Engineering. Consulting. And customized Open Source projects for companies of all sizes. Learn more about what Klarrio can do for you today.

Contact us

Behind the scenes

We created a high-level domain-specific language (DSL) workbench using Apache Flink.

This enables the bank’s low-code data scientists to easily interact with a complex streaming data processing pipeline for online payment fraud detection.

The app comes with a local continuous integration / continuous deployment (CI/CD) environment to develop, test, and analyze data in a scalable technology.

The Technology

  • Apache Flink
  • Continuous integration & deployment (CI/CD)
  • Domain-Specific Languages

The Expertise

  • Real-time analysis
  • Continuous Integration
  • Continuous Deployment

“In this project, the challenge was to allow a low-code data scientist to work on a high-performance, complex streaming data processing pipeline.

To achieve this, we abstracted the complicated plumbing (connection and joining of data streams) out of the business logic, and then we geeked out by creating a Domain Specific Language (DSL) in Scala that the data scientist could use to succinctly specify the business logic.

What did I learn? We evaluated Flink SQL as an alternative, so I learned a lot about streaming and temporal SQL.

Additionally, this was my first time working in the financial industry, seeing the security procedures in place was an eye opener.”

The fun bit for me was the whole domain specific language specification and implementation: very geeky, technical yet creative work.

Dominique ChanetLead Architect

Want to work on similar projects?

Join us!

We're hiring

Other Projects

Just a few projects examples.