At Oracle Order Management Cloud, I built a centralized recovery service for failed bulk order ingestion, including REST APIs and an LLM-based agent that helps customers remediate and resubmit failed orders. It cut manual triage by 70%.
I also re-architected an order-cleanup job on a containerized cloud platform, reducing processing time by 40%, and migrated approval rules across customer environments. Outside work, I built Order Up!, a Java job orchestration engine that ran 61,000 jobs with zero failures.

