Smart Restaurant POS
A café point-of-sale with QR ordering, a kitchen screen, and a plan for when the internet drops.
The problem
Restaurant software tends to break at the seams rather than in the middle. A second branch, a shift handover with cash to reconcile, a kitchen that can't see the queue, or a router that drops out while forty covers are seated.
What we built
We designed the system from a full spec and built three surfaces: a web POS for staff and admins, a Kitchen Display System, and a mobile ordering page customers reach by scanning the QR code on their table. Five roles run through it, from Super Admin down to Customer. Sessions open and close with cash reconciliation independent of table state, orders group by table, and the kitchen receives status changes over WebSockets. Prep times start rule-based and shift to an ML estimate as data builds up. When the network goes, the POS keeps writing locally and syncs on reconnect.
- Multi-branch setup with five levels of role-based access
- POS sessions with cash float reconciliation, independent of table status
- QR ordering at the table, grouped table orders, and a real-time kitchen display
- Auto-combo suggestions that a Branch Admin approves before they go live
- Razorpay, Stripe, cash, and UPI, with split billing
- Offline mode with local storage and automatic sync when connectivity returns
- Prep-time prediction, rule-based first and ML-refined afterwards
My contribution
What I owned on the team
Worked with the team across the POS architecture: requirements-driven design, role-based workflows, order and session flows, kitchen coordination, payment gateway integration, and the APIs behind restaurant operations.
How it turned out
Delivered as a team, from the requirements document through a working stack covering the floor, the kitchen, payments, and reporting.