Every business that moves money attracts fraud, and every fraud check has a cost: block too little and you lose money, block too much and you turn away honest customers. Fraud and risk engineering is the work of finding that balance, and doing it in the few milliseconds before a payment is approved.
You'll start with how fraud detection systems work: simple rules that catch known patterns, anomaly detection and machine learning that catch new ones, and why reducing false positives matters as much as catching fraud. Then you'll look at 3D Secure and strong customer authentication, which add a verification step for risky payments and can move the cost of fraud-related chargebacks from the merchant to the card issuer. Finally, you'll see a hands-on project that trains a model on real card transaction data.
Remember that a chargeback happens when a cardholder disputes a payment with their bank and the funds are pulled back from the merchant. Not every chargeback is fraud, since some come from confusion or merchant errors, so a good risk system tracks the reasons as well as the totals. Once you've finished this module, move on to Distributed Transactions & Messaging.
An explainer on how fraud detection works, covering rules and machine learning models, anomaly detection and behavioral analytics, and how to reduce false positives and false negatives.
How 3D Secure evolved into 3DS2 with risk-based, frictionless checks, what strong customer authentication requires in the EU, and how liability for fraud-related chargebacks shifts to the issuer.
A hands-on data science project that covers the kinds of online fraud, how to approach fraud detection with data, the challenges involved, and a demo of building a credit card fraud model. This is a long, practical video, so work through it alongside the code.