All projectsProjects / Min-Jee platform.
Full-Stack · Ecommerce · Lead

Min-Jee platform.

An ecommerce and event-booking system for a real Davao City party-needs business — replacing a fully manual operation with storefront, inventory, payments and POS. I led development across the stack.

TypeEcommerce & booking platform
StackNext.js · FastAPI · Supabase
InfraDocker · AWS Lambda · CI/CD
RoleDevelopment lead
Screenshot
Min-Jee storefront / booking flow
Fig. 01 — Min-Jee storefront / booking flow.

01The client & problem

Min-Jee organises events of every scale and owns a warehouse of rentable party supplies — and at the time was run almost entirely by hand: customer records, orders, inventory, sales and logs all manual. The brief was concrete: eliminate the manual recording, add proper payment methods, separate admin and customer accounts, and introduce point-of-sale receipts.

02My role

I led development and worked full-stack:

  • Backend endpoints covering authentication, business logic and edge-case handling, with improved error handling so failures surface cleanly.
  • Frontend modules including the rentals and point-of-sale systems, made responsive across screen sizes.
  • Integration — reconciling the frontend/backend contract drift that surfaces once the two sides actually talk, with Axios + Redux Toolkit and Zustand.
  • Containerised the FastAPI backend with Docker and built the CI/CD pipeline deploying it to AWS Lambda via ECR, automated through GitHub Actions.
Screens
Min-Jee screen 2
Min-Jee screen 3
Min-Jee screen 4
Min-Jee screen 5
Min-Jee screen 6
Min-Jee screen 7
Min-Jee screen 8

03Architecture

A Next.js frontend talking to a FastAPI backend with Supabase behind it. The read side runs through Apollo GraphQL for a typed query layer; the rest goes through Axios. The more interesting half is the delivery path — the FastAPI backend is containerised and deployed to AWS Lambda as a container image through ECR, the whole cycle automated by GitHub Actions on push. A serverless backend that scales to zero when idle, with hands-off deployment.

Where I first learned the serverless toolchain hands-on — local container to automated push-to-deploy.

— Infrastructure

04What it demonstrates

Leading a real build for an actual business against real operational problems — end-to-end ownership across backend, frontend, the GraphQL data layer and the deployment pipeline, plus the integration judgment that only comes up when you're the one connecting the pieces.

AWS Lambda
Serverless, scale-to-zero
CI/CD
Push-to-deploy
Lead
Development lead role
Next.jsFastAPISupabaseGraphQL · ApolloRedux Toolkit · ZustandDockerAWS LambdaGitHub Actions