Fraud can be crippling to a business. It hurts your revenue, reputation, and customers. Fintech fraud is a super complex space, with bad actors using a variety of attacks like identity attacks, credit card theft, and phishing scams, it’s a lot for any company to tackle on their own. Sophisticated fraudsters leverage weaknesses in protocols like SMS, the phone system, email, and DNS.

Soups Ranjan, CEO and Founder of Sardine, joins the show to discuss the different types of fintech fraud attacks that take place and how Sardine uses machine learning to automatically detect and prevent fraud.

Soups has a PhD from Rice University in denial-of-service attack prevention and has been working in fraud detection for a decade across companies like Yelp and Coinbase. With a strong background in data science and a ton of real world experience, Soups is an expert in this space.

Topics:

How did you learn fraud prevention?
What is fraud for fintechs and how is this different from other forms of fraud like ecommerce fraud?
Who are the fraudsters?
Is detecting fraud for crypto harder than other forms of fraud detection?
What were some of the tools and technologies you’ve built to help reduce fraud?
Why is machine learning the right approach? Are there ever humans in the loop as well?
What’s the input to the model?
How does training work? Where are the labels for training coming from?
What does it mean to deploy a ML model? How do you know the model is an improvement?
How has your experience in the fraud space led to the founding of Sardine
How does Sardine help optimize someone’s fraud pipeline?
What’s the typical evolution for fraud detection that a company goes through? Do they start out trying to DIY something?
What impact does Sardine have for a business and how quickly do they see ROI?
What are your thoughts on the future of fraud detection?
Are there technologies in this space that you are particularly excited about?

Resources: