Machine Learning Advanced Pro

Case Study: Real-Time Payment Fraud Detection (Stripe / PayPal-style)

Walk through a full ML system design interview answer for in-authorisation fraud scoring

30 min read 6 views

Model interview answer for real-time payment fraud detection: sub-100 ms decisions, 0.1 % positives with asymmetric costs, delayed chargeback labels, point-in-time features and velocity counters, GBDT plus graph signals, PR-AUC and dollar-weighted evaluation, cost-based decision policy and a streaming serving architecture.

Preview Subscribe to unlock

Practice questions (5)

Why PR-AUC Beats ROC-AUC at 0.1% Fraud Prevalence

Advanced · Free

View →
Label Maturity and Selective Labels

Advanced

View →
Cost-Based Thresholds Across Transaction Amounts

Advanced

View →
Training/Serving Skew in a Velocity Feature

Advanced

View →
Responding to an Adaptive Fraud Ring

Advanced

View →

Quizzes (1)

Case Study: Real-Time Payment Fraud Detection Quiz

0 attempts · 10 min · Pass 70% · Advanced

Take quiz →