Platform & InfrastructureADA Asia2022 – 2024Production

Model Metric Dashboard

Internal ML governance dashboard tracking train/test MAPE, MAE, RMSE, and classification metrics across all deployed platform models.

20+

Models Monitored

The Problem

With 20+ models in production, there was no systematic way to detect model degradation, validate retraining, or monitor performance drift over time.

What Was Built

Built an internal dashboard tracking MAPE, wMAPE, MAE, RMSE, F1, PR-AUC, and classification metrics per model per scoring run. Provides longitudinal view of model health across the full platform.

Business Impact

Enabled early detection of model degradation and provided evidence for retraining decisions across 20+ production ML models.

Tech Stack

PythonBigQueryDashboard

Domain Tags

ML GovernanceModel MonitoringMetrics TrackingMLOps

Details

Role
Primary Owner
Status
Production
Tier
Tier 3
Period
2022 – 2024
Employment
ADA Asia