Feature Engineering & Data Leakage

Leakage is one of the fastest ways to fail an MLE interview because it shows weak production intuition.

Common leakage patterns

• using future information (post-outcome fields)
• target leakage through derived aggregates
• using labels as features (directly or indirectly)
• sampling bias (logging only shown items)

Defensive checklist

• define what is available at serving time
• version features and enforce freshness
• validate training/serving parity
• monitor feature coverage and distribution

Feature ownership

Production ML improves when feature columns have owners, documentation, and monitoring.