Shreya Kamath

Built the Lane Health System, a production-grade pipeline monitoring lane health across Uber Freight 4PL accounts. Designed two independent scoring models using changepoint detection and SHAP for customer health and opportunity assessment, then engineered an 11-trigger playbook translating lane symptoms into plain-English seller actions. Validated outputs against an existing system to confirm analytical reliability and partnered across Product, Client Engagement, and Operations to define metrics and present findings to executives.

Designed and implemented real-time data pipelines on AWS using Prophecy's low-code Spark platform, reducing ID Graph processing latency by 40%. Refactored complex Spark Scala logic with AI-assisted tooling to improve maintainability while preserving functionality. Applied automated data validation and governance checks that cut pipeline failures by 80% and built a code mapping script to optimize future pipeline development.

Compared 12 LLMs via prompt engineering to improve data extraction accuracy and built 10+ web scraping solutions using Selenium and BeautifulSoup. Delivered scraping and LLM solutions 20% faster through Agile SCRUM sprints. Built Python dashboards with Matplotlib and Seaborn to visualize germination patterns, improving crop yield decisions by 15%, and optimized SQL queries for large-scale agricultural datasets.