Overview
A digital service provider in the care sector supplies people in need of care with legally guaranteed care supplies and handles the complete communication with health insurance providers. The entire process – from application submission through approval to recurring deliveries – ran on an internally developed platform that had not been systematically maintained for a long time. Organically grown structures in the data model made extensions increasingly laborious and error-prone.
Instead of spot repairs, we decided on a complete rewrite. The goal was a clean, extensible architecture that reliably supports current operations while leaving room for future features – such as new product categories or additional interfaces to health insurance providers.
The result is a modern fullstack application with a React frontend and a Django backend that noticeably simplifies the team’s day-to-day operations and lays the foundation for further growth.
My Role
As a fullstack developer, I implemented this project in solo development from January to April 2025. That meant: responsibility across the entire stack, from data modeling in PostgreSQL through API development with Django to the React frontend. Infrastructure decisions, setting up the CI/CD pipeline, and the Azure deployment were also part of my remit. Close coordination with the domain team ensured that the existing business processes were mapped correctly and optimized at the same time.
Tasks
- Redesign of the data architecture – Designing a normalized, extensible data model in PostgreSQL that cleanly maps the complex relationships between customers, health insurers, applications, and deliveries
- Development of the React frontend and Django backend – Building a TypeScript-based React application with a Django REST API as the backend, including authentication and permission management
- History feature for customer relationships – Implementation of a complete change history so the team can always trace which steps took place for a customer and when
- Handling large volumes of PDFs – Efficient processing and assignment of health insurance documents that arrive in high volume and must be reliably matched to the right cases
- Process automation – Automating recurring workflows such as status transitions, reminders, and document generation to significantly reduce manual intervention
- CI/CD pipeline and deployment – Building an end-to-end pipeline with GitHub Actions for automated tests and deployments to Azure
Technical Highlights
Clean restart instead of migration: One of the central challenges was transferring the data from the legacy system into the new model without interrupting ongoing operations. Carefully written migration scripts made the transition smooth, with data integrity ensured through extensive validation steps.
PDF processing at scale: In the care supplies sector, multiple documents accumulate per case – approvals, rejections, queries from the insurers. The system had to be able to process large volumes of PDFs performantly, assign them correctly, and make them quickly findable for the team.
Traceable customer history: The history feature goes beyond a simple change log. It maps the entire lifecycle of a customer relationship – from the initial inquiry through approval processes to recurring deliveries – giving the team a complete picture at a glance.