Built a Trip Execution microservice from a legacy monolith to support diverse logistics including P2P and third-party delivery.
•Designed integrations of new microservice to 10+ applications including trip optimizer, pricing, driver UI, and driver search system
•Created a phased rollout and ramp-up plan to de-risk migration to microservices during high-traffic sale periods.
•Incorporated multi-tenant framework with infra configurations to support scalability and new market adoption (Canada, Mexico)
Engineered a scalable machine learning model to adjust driver earnings, saving $33M annually without affecting service KPIs.
Developed a predictive driver search model, reducing offer response times and improving acceptance by over 8%.
Led the migration of Last Mile Delivery data systems from Oracle to cloud based Azure SQL with sharding implementation while maintaining sub 100ms query performance and saving infrastructure cost by 40%
•Handled processing of ~2 TB/week of structured transactional data and ~100 TB/week of event and telemetry data
•Evaluated multiple approaches incorporating using decision factors such as dev cost, deployment, infrastructure, scaling, external dependencies, disaster recovery and future risks to find the best option
Spearheaded the re-architecture of curbside pickup system, boosting adoption by ~35% with robust multi-app integration.
•Built a cloud-native, auto-scaling geofencing service supporting complex shapes, cutting wait times for drivers by 40%.
•Developed UI and backend for geofence CRUD operations, enabling store-level detection for optimized pickup logistics.