# Hospital Inventory Management System

URL: https://mepritam.dev/work/hospital-inventory-management-system/

Concurrent inventory management services in Go & Gin with PostgreSQL, JWT role-based access control, near-expiry alerts, and Docker health monitoring.

---

### High-Concurrency Inventory Control for Clinical Operations

In healthcare facilities, stock discrepancy or expired inventory is not merely a financial loss; it directly impacts patient care workflows. I built a resilient, concurrent backend service in **Go (Golang)** and the **Gin web framework**, giving hospital pharmacists and clinical teams real-time visibility into pharmaceutical stock, batch expiry timelines, and audit records.

#### Architectural Highlights & Implementation

- **Concurrent API Performance in Go:** Leveraged Go goroutines and Gin routing to support 50+ concurrent medical staff members executing real-time barcode scans, check-ins, and batch updates without lock contention.
- **Role-Based Access Control (RBAC) & JWT Security:** Implemented cryptographic JWT authentication middleware enforcing granular role permissions (Super Admin, Pharmacist, Nurse, Auditor) across all API endpoints to prevent unauthorized stock alterations.
- **Near-Expiry Alerting Engine:** Designed background Go worker routines querying **PostgreSQL/GORM** with composite indexes (`expiry_date`, `batch_id`, `category`), generating automated near-expiry alerts and Excel reports. This cut reporting turnaround by **40%** and reduced medication expiry waste by **20%**.
- **Operational Resilience & Containerization:** Structured the backend within lightweight multi-stage **Docker** containers with automated health checks (`/healthz`) and restart policies that guaranteed **99.9% uptime** across hospital operational shifts.

#### Key Engineering Decisions

1. **Why Go & Gin:** Selected Go over Node.js for this service due to strict memory footprint constraints on on-premise hospital gateway servers and Go's static typing, ensuring predictable p99 latencies under intake surges.
2. **GORM Connection Pooling:** Configured aggressive PostgreSQL connection pooling (`SetMaxIdleConns`, `SetMaxOpenConns`, `SetConnMaxLifetime`) to prevent socket exhaustion during peak outpatient shifts.
3. **Structured Audit Logging:** The service records every inventory modification as an immutable transaction log, providing an auditable trail required for healthcare regulatory compliance.