← All work

BellyBioMe

The team capstone for the Per Scholas AWS re/Start program: a nutrition tracker with barcode scanning, food search, daily goals, and dashboards.

In short: Scan or search → Match USDA data → Log the meal → Track goals

Why I built it

Logging meals by hand is slow, and nutrition labels are hard to compare or turn into daily goals.

What exists

The Django app was deployed on AWS Elastic Beanstalk during the program and is offline now. Its code covers meal and gut-reaction logging, barcode scanning and image upload, similar-food search over USDA data, daily nutrition goals, and progress dashboards.

How it works

A Django app on PostgreSQL. USDA branded-food data in DynamoDB backs barcode matching, a vector index over USDA SR Legacy and Foundation data backs similar-food search, S3 holds images, and an API Gateway endpoint handles image cleanup.

  • Meal and gut-reaction logging
  • Barcode scanning and image upload
  • Similar-food search over USDA data
  • Daily nutrition goals and dashboards

My role

Developed and deployed the application as the team's software developer: the Django app, barcode and vector-search features, the AWS integration, and the Elastic Beanstalk deployment.

What I checked

  • deploymentDeployed on AWS Elastic Beanstalk during the program
  • repository219 commits, July to October 2024

What stays bounded

  • Not a medical or dietary tool
  • Offline; there is no live demo

What is missing

  • The planned serverless image pipeline was never finished.

Next

None; kept as earlier work.