Client name kept private. Same retail group as the card payments project.
The problem
A retail group in Saudi Arabia, with more than 110 stores and over 20 brands across several cities, was rolling out RFID tags on its products. RFID only pays off if every scan flows into the systems the stores already run: the ICG point-of-sale (POS) system for sales, and inventory for stock. Out of the box, the two didn't talk to each other.
What I built
- POS ↔ RFID backend. A custom backend that links the group's ICG point-of-sale systems with the RFID platform's APIs, so RFID data is used for sales, item tracking and inventory.
- EPC encoding and decoding. Converting between the EPC values stored on RFID tags and the group's own product codes, so a scanned tag resolves to the right item.
- Inventory synchronisation. Python APIs that keep stock in sync between the RFID platform, the POS and the group's enterprise systems.
- Repeatable store rollout. Start-up scripts and an installation guide, so the same setup can be installed at each store and city.
The result
Item scanning became 30% faster, across a rollout covering 110+ stores and 20+ brands.
Stack
Python · RFID / EPC · ICG point-of-sale · REST APIs · enterprise system integration
Related work for retail
On a separate contract I built a Django and React dashboard that unified sales, footfall and store data from several sources for multi-brand retailers, with Python pipelines that ingest, clean and transform large Microsoft SQL Server datasets for reporting.
What this means for your project
Hardware and legacy-system integrations reward careful data mapping more than clever code. If you need devices, an existing POS or ERP and a modern dashboard to agree with each other, I've done it in production at scale.