Yu Zhao

● Developed a Claude AI-driven system to classify proof-of-delivery images against shipment metadata, optimizing for high precision to ensure auto-approved PODs are reliably correct and routing uncertain cases to human review; automated the process with Playwright, cutting processing time to 8 seconds per image. ● Reduced warehouse labor costs by $1,360 per day by using time series models to forecast inventory and optimize sorting staff allocation. ● Built an automated ETL pipeline in Python to integrate SQL databases and internal operational systems, enhancing the workflow with Playwright and Claude Code into a lightweight ERP data agent supporting daily operations across five service sites, reducing data processing time by 68%. ● Delivered 7+ daily data analysis reports (package returns, weekly payment, warehouse labor efficiency/inventory, mis-sorts, backlog reports, and daily KPIs) and conducted ad-hoc root cause analysis on KPI declines to support decision-making.

● Consolidated 240+ SharePoint Excel files and 6+ SQL databases into a Databricks Delta Lake, enabling scalable analytics and automated reporting. ● Diagnosed and resolved a critical dashboard failure impacting 32,000+ users within 2 hours, identifying root cause (data type drift) and rebuilding pipelines using PySpark. ● Optimized cross-time-zone data pipelines (10+ regions), reducing runtime by 22% and improving data consistency. ● Developed client-facing Power BI dashboards visualizing real-time vehicle performance (fuel efficiency, temperature alerts, geospatial tracking).

● Increased sales revenue by $14,000 by identifying new demand patterns through analysis of 300+ customer surveys. ● Built dashboards and visualizations for monthly and quarterly business performance reporting. ● Developed Python pipelines to clean and transform Google Sheets data into MySQL, enabling automated reporting and statistical analysis.