The Problem
A private cultural platform needed merch e-commerce to feel simple for customers and operationally reliable for staff. The tricky part was not just showing products: the system had to coordinate variants, stock, payment checkout, order records, image handling, shipping status updates, and customer notifications without exposing private customer or payment data.
This case study keeps the platform name, repository, production endpoints, private product media, customer details, and order records unpublished.
What I Built
I treated e-commerce as its own product slice instead of folding it into a generic platform story. The source shows a full customer path from shop browsing to cart state and checkout, plus admin surfaces for products, variants, orders, uploads, shipping/tracking updates, and product statistics.
Project Overview
This is a private client platform. The public case study uses source-backed architecture summaries and recreated visuals only.
My Contribution
- Built the merch e-commerce data path around products, variants, cart state, checkout creation, orders, and order items.
- Modeled variant-level details, reserved-stock fields, order status, third-party carrier shipping, tracking metadata, and fulfillment-ready identifiers without publishing real order data.
- Connected admin product management, image uploads, customer order notifications, shipping/delivery notifications, and Stripe-backed checkout/webhook behavior.
E-Commerce Flow
Merch transaction flow
The public storefront and admin console share durable product, variant, order, and inventory state through the API.
Diagram uses generic labels and omits product names, customer data, payment IDs, URLs, and private media.
Technical Decisions
Coordinate checkout reservations and order finalization behind the API boundary.
- Merch can involve simple products, size/color variants, temporary checkout reservations, payment confirmation, and shipping operations.
- The checkout and webhook paths have to keep inventory, order status, and reservation cleanup in sync.
- Customer-facing checkout stays simple while durable order and inventory state remains controlled by server-side payment events.
How I Work
Engineering Challenge
- E-commerce case studies can easily overclaim by leaning on revenue or order volume.
- I traced only source-backed objects and flows: product records, variants, cart operations, checkout route, order models, admin product pages, shipping/tracking updates, upload handling, and order emails.
- I published the architecture and workflow shape, but withheld product media, customer/order records, payment IDs, private URLs, and volume metrics.
- The story shows full-stack e-commerce depth while staying privacy-safe.
Quality & Testing
Quality Evidence
Product engineering evidence
- Private source includes e-commerce models, migrations, frontend API clients, cart state, admin product pages, order pages, shipping/tracking updates, and webhook handling. No public test counts are claimed.
- Private deployment evidence includes Dockerized API structure and CI/CD service shape, but no private run logs or deployment identifiers are published.
- Source evidence verifies Next.js, React, TypeScript, FastAPI, PostgreSQL, SQLAlchemy, Alembic, Stripe checkout/webhooks, image upload, and email notifications.
Public page quality
- The public visual is a recreated SVG with accessible title and description.
- The diagram is scalable inside the shared case-study layout.
- No repository URL, production endpoint, payment record, customer record, or private product image is bundled.
| Area | Evidence | Published detail |
|---|---|---|
| Storefront | Frontend shop route, cart store, merch API client | Catalog-to-checkout workflow |
| Backend | Product, variant, order, webhook, upload, and email source areas | Generic e-commerce architecture |
| Admin | Product, variant, order, and product-stat screens | Staff operations scope |
Results & Impact
This slice demonstrates full-stack e-commerce engineering: customer browsing, cart state, payment checkout, durable order records, variant inventory, admin product management, image upload, admin-managed third-party carrier shipping, tracking metadata, order status updates, and customer messaging. The public version intentionally avoids revenue, traffic, automated fulfillment-handoff, and production-availability claims.
What I Learned
Private e-commerce work still needs a precise story. The strongest publishable version is not "I built a store"; it is the system boundary between customer intent, payment confirmation, inventory state, shipping operations, and staff workflows.