
PixFood is an AI-driven food app that uses advanced image recognition to deliver a hyper-personalized experience to users through custom-tailored recipes.
PixFood was a crucial block in IMP’s (International Masters Publishers) digital transformation journey, giving them a competitive edge in this highly saturated landscape.
Problem Definition
PixFood wanted to be able to include videos, images and gifs in the app and to also be able to stream them in an optimised way according to their quality and the internet speed.

Proposed Solution
To solve all problems PixFood had, we decided to use AWS services.
First of all, we used Lambda as a serverless computer, S3 hosting as an admin panel for static sites, and then CMS and Storage Asset for customer engagement campaigns (such as emails, push notifications, in-app messages etc.).
Also, we were able to compress the videos when uploading, thus creating more different formats for each video. That way, users could choose a suitable format at a specific timing.
Additionally, we used the Simple Email Service for enabling our client to send emails, and its users to be able to activate their email addresses or change password.
All data from the image recognition feature was stored with Dynamo DB and further used for analysis and user recommendations.
We also used Lambda and created a cron job that does regular checkups every 24 hours. If there is a new user subscription or a premium subscription payment, there is a call made through Lambda. That way, we forward all information to AppStore or Playstore for verification. After the verification is finished, the user receives their premium subscription.
Outcomes of the Project

The ultimate benefit for the users from compressing the videos was a more effective app running and video loading regardless of users’ internet connection. Hence, there was no lagging, no buffering, and no excessive internet usage.
With the deployment of all above-mentioned AWS services, we brought the app to a level where it can easily serve 10K or 10M users. The app became very dynamic – we could easily add or remove features.
The cost was also flexible – meaning we were able to automatically adjust to the product needs. The app was fast-built on AWS and hosted close to the user usage and it has zero downtime.