This application had a simple premise - Put a bottle of a beverage in front of the camera and the user will get all the info about the beverage. This application implemented Google Cloud Vision, Machine Learning, Data Science and Hybrid App Development. It employed a user trained model which has trained by the team with almost a TB of scraped data from the internet. Initially, it was aimed towards hard beverages but the client had plans to make it work for any type of beverage in the future.
Training the ML model with more than a TB of scraped data was a challenging task. Another challenge was working with React Native and Google Cloud Vision API since React Native is ever evolving and things keep changing pretty fast so the team was looking for all the updates in the framework that could affect the application.
The team selected the ML model after a lot of research which allowed faster responses from the Google Cloud Vision API and thus the application being more responsive to the user.
This application is yet to be launched on the Google Play Store and Apple App Store but has received great reviews from users which have used the early prototypes. Various liquor store owners have also shown interest in implementing this application in their store premises.