Recovery Management System (RMS)

A full-stack application for coordinating roadside recovery and vehicle breakdown services, serving as a central hub that integrates multiple external systems.

Key achievements include implementing bi-directional SOAP-based synchronization with the external system, REST API integration with external provider for breakdown service management, and SFTP-based fleet synchronization with one of the biggest fleet companies in the UK. The system features real-time job tracking, comprehensive error logging with notifications, and a dynamic feature flag system for flexible deployments. My role involved architecting the entire monorepo structure, implementing queue-based async processing with retry mechanisms, building data export functionality, and ensuring reliable integration between legacy SOAP systems and modern REST APIs.

The monorepo holds the NestJS API and the React dashboard over one Prisma schema on PostgreSQL, and most of that schema exists to reconcile systems that disagree. A WSDL-defined SOAP service, a REST booking provider and a scheduled SFTP feed share no identifiers, do not agree on when a record is final, and any of them can be unavailable when we need them. So the design leans on durable state rather than on calls succeeding: a deferred path for outbound updates that cannot be sent yet, snapshots of inbound external payloads, full change history per job, and separate error and request log tables so a failed exchange can be replayed rather than guessed at.