Sasuai App v2 (Hybrid Distributed Enterprise POS)

1. Problem

Retail chains required centralized, cloud-based analytics to manage multiple stores simultaneously. However, purely cloud-based POS systems failed at the store level due to slow internet, printer driver issues, and hardware integration limitations.

2. Solution

Architected a hybrid distributed system. I split the application into two distinct layers: a Next.js cloud backend for global analytics and Admin controls, and a lightweight Electron desktop app for the cashiers to ensure hardware compatibility (thermal printers/scanners) and offline resilience.

3. Architecture

• Global Admin/Backend: Next.js + Node.js
• Store Terminal: Electron + React
• Database: PostgreSQL (Cloud)
• Message Broker: RabbitMQ
• External Integrations: Blastify API (WhatsApp Engine)

4. Key Engineering Decisions

• Decoupled Cloud/Desktop Architecture: Electron handles local hardware bridging (USB printers) and fast checkout UIs, while Next.js handles heavy data aggregation and multi-store management.
• Event-Driven CRM Notifications: Integrated RabbitMQ to handle WhatsApp membership notifications asynchronously. When a customer reaches a new loyalty tier, the POS fires an event to the queue, and a backend worker triggers the Blastify API to send the WhatsApp message, ensuring the cashier's UI is never blocked by network latency.

5. Challenges

• Architecting robust data synchronization between the offline-capable Electron terminals and the central PostgreSQL cloud database.
• Managing RabbitMQ queue failures and implementing retry-mechanisms for WhatsApp notifications if the Blastify API experienced downtime.

6. Result

• Delivered an enterprise-grade POS ecosystem capable of managing multi-store retail chains.
• Achieved seamless hardware integration at the local level without sacrificing global cloud analytics.
• Increased customer retention through automated, real-time WhatsApp loyalty notifications.

7. Future Improvements

• Implement full offline-first Conflict-Free Replicated Data Types (CRDTs) using local IndexedDB to allow stores to operate for days without internet and sync perfectly upon reconnection.
• Introduce containerized deployment (Docker/Kubernetes) for the Next.js backend to support massive horizontal scaling as new stores are onboarded.