PharmaWMS Enterprise
GSP-compliant Pharmaceutical Warehouse Management System featuring automated FEFO (First-Expired, First-Out) batch allocation, cold-chain monitoring, and LightGBM AI demand forecasting.
Lead Architect
Nguyễn Đăng Khoa
Architecture
Spring Boot 4.1 Modulith
Domain
Pharma Cold-Chain & AI
Timeline
2025 - 2026
Engineering Challenge
Pharmaceutical warehouses operate under strict Good Storage Practice (GSP) regulations where expired batches, cold-chain excursions, or race-condition stock discrepancies can compromise patient safety.
Traditional inventory systems rely on simple FIFO rules and manual stock forecasts, leading to costly pharmaceutical spoilage and stockouts of critical medicines. The challenge was to engineer a transactional, concurrency-safe FEFO (First-Expired, First-Out) allocation engine paired with an AI microservice capable of predicting multi-horizon demand spikes.
System Architecture & AI Solution
I architected PharmaWMS using Spring Boot 4.1 Modulith with strict domain-driven bounded contexts (Inventory, Inbound/Outbound Orders, Cold-Chain Compliance, and Audit Logging) backed by PostgreSQL with pessimistic/optimistic locking for zero-race-condition FEFO batch allocation.
Alongside the core Java backend, I built a dedicated Python FastAPI + LightGBM AI microservice that analyzes historical dispensing velocity, seasonal disease trends, and batch expiry windows to generate automated reorder alerts—surfaced in real time on a responsive Next.js 16 operator dashboard.
Key Technical
Highlights
• 100% GSP-compliant FEFO (First-Expired, First-Out) automated batch picking with real-time quarantine locking.
• Decoupled Spring Modulith event-driven architecture with full CI/CD verification on GitHub Actions.
• Fast LightGBM demand forecasting API via FastAPI with an interactive Next.js 16 dashboard.
Nguyễn Đăng Khoa (@minKasent)
Source Code & Documentation on GitHub (minKasent/pharmawms)
Let's build something extraordinary!
