In Progress
Virtualy
AI voice-coaching platform for realistic roleplay practice
ReactTypeScriptJavaScript+9
CompletedE-Commerce
FBM Insight — Amazon FBM Order Management System
Turning hours of manual order matching into seconds — for Amazon FBM sellers.
A full-stack order management platform built for Amazon FBM (Fulfilled by Merchant) sellers who dropship through Essendant. FBM Insight automates order matching between Amazon and supplier data, provides real-time profit analytics per store, and supports multi-store isolation — replacing hours of manual spreadsheet work with a single, streamlined dashboard.
Role: Full-Stack Developer
Client: Jamilur Rahman
Team: Solo developer
Started: October 2025
Completed: February 2026
98.2%
Match Accuracy
3x faster than manual
Processing Speed
~10s (Turbopack)
Build Time
22 endpoints
API Routes
10 app routes
Pages
The Challenge
Amazon FBM dropshippers managing multiple stores were spending hours daily in spreadsheets — manually matching Amazon orders with Essendant supplier invoices by PO number, calculating profit margins per SKU, and tracking refunds across stores. A single mismatch could mean undetected losses running for weeks. There was no unified view of profitability across stores, and CSV imports from Amazon often broke due to product names containing inch-mark quotes (e.g. 24") that corrupted parsers.
The Approach
Built a modern Next.js 16 application leveraging Cache Components and Partial Prerendering for instant page loads. Designed a smart matching engine that automatically pairs Amazon orders with Essendant supplier data by PO number. Implemented store isolation via URL parameters so users can seamlessly switch between stores while preserving navigation context. Used Server-Sent Events for streaming CSV imports to handle large files without timeouts on Vercel's Hobby plan. Built a companion Chrome Extension (Manifest V3) to scrape Essendant orders directly from the supplier portal and sync them into the dashboard with deduplication.
Key Features
Automatic order matching by PO number
Multi-store management with store isolation
Real-time profit & loss analytics per store
CSV import with SSE streaming progress
Refund tracking and margin alerts
Chrome Extension for Essendant order scraping
Store-level profit CSV report export
Manual order entry form
Match queue with exception handling
SKU-level performance breakdown
Responsive mobile-first dashboard
JWT-based authentication
Scope:10 app routes, 22 API endpoints, 1 Chrome Extension, Prisma schema with 7+ models, deployed on Vercel with Neon PostgreSQL
Tech Stack
Built With
Next.js 16React 19TypeScriptTailwind CSSPrisma ORMPostgreSQLNextAuth.js v5Neon DatabaseVercelTurbopackServer-Sent EventsChrome Extension (Manifest V3)
Challenges & Solutions
Challenge
Vercel Hobby plan 1MB Edge Function limit
Solution
Split auth into edge-safe config (auth.config.ts) and full Prisma-based auth (auth.ts) — middleware only imports the lightweight config
Challenge
CSV imports timing out on large files
Solution
Implemented Server-Sent Events streaming with chunked processing so uploads never hit serverless function time limits
Challenge
CSV parse errors from inch-mark quotes in product names
Solution
Pre-processing regex to replace measurement quotes (24") with safe alternatives before parsing
Challenge
Next.js 16 Cache Components + dynamic data
Solution
Wrapped dynamic client components in Suspense boundaries and used Partial Prerendering for instant static shells with streamed dynamic content
Gallery
Next.js 16 (Latest)Partial PrerenderingCache ComponentsMobile ResponsiveEdge-Optimized AuthTypeScriptAccessibleSecureDockerHigh PerformanceSEO OptimizedCI/CDResponsiveTested
Explore This Project
Related Projects
Completed
CakeOrFake — AI-Powered Upwork Applicant Screening Chrome Extension
AI-powered Chrome Extension that turns hours of Upwork applicant screening into minutes
Chrome Extension (Man…JavaScriptFirebase Cloud Functi…+10
Completed
Full-Stack E-commerce & Booking Platform
HennaHub is a full-stack e-commerce and service booking platform built for Bangladesh's henna art market. It enables a henna business to sell products online, accept service bookings, manage artists, and engage customers through an AI-powered chatbot — all from a single unified platform
Next.js 16HonoCloudflare Workers+5