Build a fraud detection model
Company: Shopify
Role: Machine Learning Engineer
Category: Machine Learning
Difficulty: medium
Interview Round: Technical Screen
Design a machine learning approach for detecting fraudulent transactions or user actions.
Discuss:
- How to define the prediction target and labels
- What features you would build
- Which model families you would consider
- How to handle severe class imbalance and label delay
- What offline and online evaluation metrics you would use
- How to choose thresholds for actioning decisions
- How to monitor the model after deployment
Assume the business cares about both loss prevention and minimizing false positives that hurt legitimate users.
Quick Answer: This question evaluates competency in designing end-to-end fraud detection machine learning systems, including defining prediction targets and labels, feature engineering, model family selection, handling severe class imbalance and label delay, evaluation metrics, thresholding for action, and post-deployment monitoring.
