Whenever you're ready, walk me through how you'd approach this.
How would you design a dashboard to monitor model drift in production? Treat this as a hands-on pipeline design task. Explain the data sources, batch or streaming ingestion, drift calculations, storage model, orchestration, alerting, and how the dashboard distinguishes data drift, prediction drift, and performance degradation. Address delayed labels, missing data, schema changes, reproducibility, and backfills without assuming a particular business domain or model type.