About the client
The client works in a regulated national pharmaceutical pricing market. Its business covers drug-pricing intelligence, monitoring of the official pricing datasets, compliance tracking and competitive pricing analysis. Pricing data feeds market-access decisions, so the team needs to see regulated price movements accurately and quickly. The pricing authority publishes its data as large Excel files, release after release.
Finding price changes without opening a spreadsheet
- Large daily files. Datasets had to be ingested every day without timeouts or memory failures, which ruled out loading whole files into memory.
- Reliable change detection. Changes to pharmacy purchase and retail prices had to be detected accurately, and every past value had to remain queryable.
- Data-layer security. Authentication, role-based authorisation and database row-level policies had to work together so each user saw only the records their role allowed.
- Dependable scheduling. Ingestion and alerting had to run every day, and changes to serverless functions carried regression risk across authentication and processing.
- Privacy. Logging had to be GDPR-compliant, with personally identifiable information protected throughout.
What FiftyFive built
FiftyFive delivered the platform from architecture to production deployment: development, QA, security testing and release. The team included an AI-enabled full-stack developer who used AI-assisted development tooling to speed up the build. The scope covered an automated ingestion pipeline, price-change detection with historical versioning, a secured PostgreSQL data layer, scheduled email alerting and a React web interface.
How the solution works
Daily ingestion
- Scheduled cron jobs start ingestion of the latest pricing release every day.
- Excel files are read with streaming and chunking, so large datasets load without exhausting memory.
- Supabase Edge Functions run the event-driven processing steps.
Change detection and price history
- Each new release is compared with stored values.
- Triggers fire when a pharmacy purchase price or retail price changes.
- Prices are versioned rather than overwritten, so the full history across the dataset stays available for analysis.
Security in the data layer
- Auth0 handles authentication and role-based user management.
- PostgreSQL schemas use indexing, Row-Level Security policies and access controls, so authorisation is enforced in the database, not only in the interface.
- Logging is GDPR-compliant, personal data is protected, and endpoints were hardened for production.
Alerts and interface
- Scheduled cron jobs send email alerts through the Resend API, as transactional messages and digests, tied directly to detected price changes.
- Users work in a React 18 and TypeScript interface built with Vite, Tailwind CSS and shadcn/ui, deployed on Vercel.
Technology stack
What the solution enabled
- Data Engineering streaming ingestion, change detection and price versioning for large regulated datasets.
- AI-Enabled Engineers an AI-enabled full-stack developer using AI-assisted tooling to deliver a production platform in two months.
- Web application development a React and Supabase application with database-enforced security.
- Software testing and QA QA and security testing as part of delivery.
Business transformation
To a production drug-pricing platform
Automated ingestion of the official pricing datasets
Detection of pharmacy purchase and retail price changes
A production platform in two months
Covering architecture, development, QA, security testing and deployment.
Daily automated ingestion
Of the official pricing datasets, replacing manual Excel reviews and comparisons.
Automatic detection
Of pharmacy purchase and retail price changes, with a versioned history of past values.
Scheduled email alerts
To pricing and market-access teams when prices change.
FiftyFive Technologies built a production drug-pricing intelligence platform in two months for a European pharmaceutical pricing and market-access company. The client had been comparing the national pricing authority's Excel releases by hand to find price changes. The platform ingests each release daily using streaming and chunking, detects changes to pharmacy purchase and retail prices, keeps a versioned price history, and emails scheduled alerts. It runs on React, Supabase and PostgreSQL with Row-Level Security, Auth0 authentication and Vercel hosting.
Technical questions
How do you ingest large pricing spreadsheets every day without memory failures?
In this platform, FiftyFive Technologies reads each Excel release with streaming and chunking instead of loading the whole file into memory. Scheduled cron jobs start ingestion daily, and Supabase Edge Functions run the processing steps, so large pharmaceutical pricing datasets load reliably.
How can a pricing platform restrict which records each user sees?
FiftyFive Technologies combined Auth0 role-based user management with PostgreSQL Row-Level Security policies. Access rules are enforced in the database itself, so a user can only query the records their role permits, whichever part of the application makes the request.