📘

Credit Card Fraud Detection System

Design a real-time fraud detection system for credit card transactions.

Scenario: A payment processor handles 10,000 transactions per second. Fraudulent transactions are less than 0.1% of all transactions. The system must flag fraud within 100ms while minimizing false positives (blocking legitimate purchases).

Your Task: Design the fraud detection pipeline.

Key Challenges:

Extreme class imbalance (0.1% fraud rate)
Real-time latency requirements (< 100ms)
Evolving fraud patterns
Minimizing customer friction from false positives

Design Mode

📝 Your Design Approach

Describe your system design approach. Consider components, data flow, and key decisions.

🎯 Design Questions(Select all that apply)

Q1.Why is real-time feature computation critical for fraud detection?

Q2.How should the system handle the extreme class imbalance (~0.1% fraud rate)?

Q3.What latency requirement is appropriate for transaction scoring?

Q4.How should the model adapt to new fraud patterns?

0 of 4 questions answered

📘