Z-Laundry System
Enterprise-grade on-demand laundry platform — 3 mobile apps, admin dashboard, payment processing, geospatial delivery zones, and real-time analytics serving thousands of customers across 5 years of continuous development.
The Problem
What needed to be solved
The platform started as a basic ordering app, but as the business scaled, every department needed software — operations needed delivery optimization, finance needed real-time analytics, quality teams needed inspection workflows, and marketing needed customer segmentation. The backend had to evolve from simple CRUD into a production system handling complex order pipelines, multi-provider coordination, and thousands of concurrent customers — all without downtime.
What I Built
My technical contributions
Scalable Service Architecture
Designed 72 dedicated service classes following single-responsibility principles — separating order creation, status management, pricing calculations, and query logic into focused modules. This enabled parallel feature development and reduced regression risk across a 100K+ LOC codebase.
Event-Driven Order Pipeline
Built an event-driven order lifecycle with 10+ states (created → assigned → picked → processed → delivered), supporting order splitting, partial deliveries, returns, and hard-cancel logic with provider capacity validation — all tracked with immutable activity logs.
Geospatial Delivery System
Implemented polygon-based zone matching using MySQL spatial functions for delivery coverage validation, automatic customer zone assignment, and area-split migrations — enabling the business to expand to new regions without code changes.
Performance & Caching Strategy
Introduced tag-based cache invalidation, eliminated N+1 queries with eager loading, built denormalized analytics snapshot tables for instant dashboard loads, and implemented server-side DataTables pagination for datasets of 2,800+ records.
Scalable Notification Engine
Architected a queue-based notification system using batch jobs (100 customers per batch with staggered dispatch) supporting Firebase push, Telegram alerts, scheduled delivery, per-customer delivery tracking, and retry logic with exponential backoff — processing 3,000+ notifications in under 8 minutes.
Full Codebase Refactor (2025–2026)
Led a large-scale refactor spanning 99+ files with +129K lines added and -45K lines removed — restructuring the entire backend architecture, modularizing inline assets, adding API documentation, implementing area-split logic, and improving customer filtering. The result: a cleaner, scalable codebase ready for team growth.
System Architecture
How it's built
Service Layer Pattern
72 dedicated service classes with single-responsibility, enabling parallel feature development
Event-Driven Architecture
Order lifecycle events trigger notifications, analytics snapshots, and cross-module updates
Queue-Based Processing
Batch notification jobs processing 3,000+ customers with staggered dispatch and retry logic
Repository Pattern
Data access abstraction with Eloquent scopes and query builders for complex filtering
Observer Pattern
Model observers for lifecycle hooks — auto-creating analytics snapshots on order changes
Strategy Pattern
Different calculation strategies for pricing by context (dimensions, weight, flat rate)
Scope of Work
Key features delivered
Visual Proof
App screenshots
Results & Impact