Detect Card Fraud with Imbalanced Data | Dataford Interview Questions

Welcome to your interview.

The question is on your right: Detect Card Fraud with Imbalanced Data. Take a moment with it first.

Talk your thinking through with me if you like - when you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes). Discussion and graded submissions share your five interviewer interactions, so spend them well.

Business Context

PayFlow processes roughly 8 million card transactions per day. Fraud is rare but costly, and the risk team needs a model that flags suspicious transactions without overwhelming manual reviewers with false positives.

Dataset

You are given a historical transaction dataset for binary classification.

Feature Group	Count	Examples
Transaction attributes	12	amount, currency, merchant_category, card_present, payment_channel
Customer behavior	10	avg_txn_7d, txn_count_24h, chargebacks_90d, account_age_days
Merchant signals	6	merchant_risk_score, country, device_consistency, velocity_score
Device / location	8	device_id_hash, ip_country, distance_from_home, browser_type
Temporal features	4	hour_of_day, day_of_week, is_holiday, seconds_since_last_txn

Size: 1.2M transactions, 40 features
Target: is_fraud (1 = fraudulent, 0 = legitimate)
Class balance: 0.7% fraud, 99.3% non-fraud
Missing data: ~12% missing in device fields, ~4% missing in merchant metadata, sparse missingness elsewhere

Success Criteria

A good solution should improve fraud capture materially over a majority-class baseline. Target at least 75% recall on fraud while keeping precision above 15% for the review queue, and achieve PR AUC above 0.30 on the holdout set.

Constraints

Batch scoring every 15 minutes; average inference latency should stay under 50 ms per transaction.
The fraud operations team needs feature-level explanations for flagged transactions.
False positives are expensive because analysts can review only ~5,000 transactions per day.

Deliverables

Build a classification pipeline that handles the severe class imbalance correctly.
Explain how you would compare class weighting, resampling, and threshold tuning.
Choose evaluation metrics appropriate for rare-event detection and justify them.
Produce a validation strategy that avoids leakage from customer or time-based patterns.
Recommend a production-ready model, decision threshold, and monitoring plan.