Elevator & Escalator Predictive Maintenance

Broad Category: Tabulated Analytics + Time Series

Role: Principal Data Scientist

Duration: Jan 2018 – Jan 2021

Overview

Built a predictive maintenance system for elevators and escalators, delivering failure forecasts and alerting maintenance teams across three countries.

Key Contributions

• Engineered daily elevator-level features from operational and maintenance datasets.
• Built supervised learning models to predict failures seven days ahead.
• Integrated alerts and dashboards for maintenance personnel via TKE web services.
• Scaled the platform from 10K to 200K elevators with modernized pipelines and DevOps automation.

Technology

Azure Databricks, Data Lake, ADF, Python, PySpark, Tableau, Power BI, Azure CI/CD.

Outcome

Delivered a large-scale production system that minimized unscheduled maintenance and improved operational reliability.