Part 1 — Verify Transaction Data Integrity

This section focuses on building the foundational data-integrity checks for fraud detection. The prompt is very long, but the logic itself is simple. We read six fields from a CSV file and verify that all fields are non-empty.

Part 2 — High-Risk Rule Validation

On top of the basic transaction information, this part introduces two core risk-control rules:

Amount rule: the transaction amount must fall within the business-defined normal range.

Payment-method rule: the payment method must not appear in the blocked method list.

If a transaction violates any of these conditions, it is flagged as SUSPICIOUS.

Part 3 — User Behavior Matching

This section requires validating whether the transaction aligns with the user’s behavioral baseline. We ensure that at least 50% of the behavioral attributes—such as commonly used spending countries, typical time ranges, and average transaction-amount intervals—match the user’s historical behavior.

Implementation requires feature extraction, normalization, and computing a match ratio.
If the current transaction’s match ratio is below 50%, it is classified as a behavior mismatch and flagged as SUSPICIOUS.

Part 4 — Smart Fraud Error Reporting

The final section implements smart error reporting, replacing the vague label “SUSPICIOUS” with specific error codes. The system outputs up to two error codes according to priority. If there are no issues, it outputs “OK”.
The output report must maintain column alignment for readability.
If you need more interview questions related to Stripe, you can refer to my leetcode homepage