Food Ordering App For Restaurants Network

Team

2 Backend engineers
1 Front-end engineer
1 Machine learning engineer
1 QA
1 PM
1 UX/UI

Timeline

9 months (in progress)

Location

Europe, Middle East

Tech Stack

Python, Django, TensorFlow, Django Rest Framework.,JavaScript, React Native, Keras, C++, PostgreSQL

Challenge

Our client wanted to create an ordering application that would cover 3 major scenarios: in-venue order, preorder, and delivery. Of course, the system should also be fully integrated with the client’s POS system in order to get the order receipt confirmation, fetch and keep synchronized menu data in order to optimize the management effort.

The client granted us quite a lot of freedom not only in technological decisions but also in the overall project approach and management. There were two major goals for us to keep in mind: provide an opportunity for upselling and cut down the operating costs and stuff routine for the venues’ administration.

Solution

We’ve started with very thorough and careful business analysis in order to get a detailed understanding of all the operations in the customers’ venues. We also interviewed local managers and analyzed current POS and its integration possibilities. That allowed us to build a very accurate flow of the application that did not only eliminate the extra load on the staff but also unleashed quite a number of resources which had a serious impact on the overall restaurant ROI.

The system comprises React Native app for mobile and Android, Web interface for system management and system backend, and is integrated with payment gateway and POS system. We used the Django framework for the web interface in which the system administrator can add new venues to the system, manage venue details, view client’s info, and orders, send promo messages and view analytics. The usage of Django standard mechanisms helped us to save a lot of resources so we could better focus on the mobile user interface.

The most exciting thing about this project was that, by coincidence, at that time our R&D department has been working on extracting value from the datasets that are very similar to those the customer could build with our system. So we offered building a personalized suggestions feature that learns frequent users’ tastes and offers personalized additions. We took our model as a basis and customized it according to the customer needs. The smart suggestions machine learning model proved to be able to upsell up to 20% for the frequent user’s segment