A forecasting and optimization stack built for aftermarket demand planning, pricing analysis, and unscheduled maintenance signal detection.

What I built

  • Multivariate time-series forecasting models using utilization, maintenance, repair, and supplier signals
  • Price elasticity models for demand-sensitive pricing analysis
  • Failure and anomaly workflows to anticipate unscheduled demand patterns
  • Reusable Databricks, Unity Catalog, MLflow, and PySpark pipelines for model delivery

Outcome

  • Turned scattered operational data into reusable decision-support assets
  • Enabled repeatable experimentation, tracking, and deployment workflows in a shared platform
  • Supported work from solution architecture through proof of concept and MVP delivery

Stack

Databricks, MLflow, PySpark, forecasting, pricing analytics, anomaly detection, and feature engineering