ParseMD – Async Referral Processing API

A production-ready backend for asynchronous medical referral processing. Converts uploaded PDFs into structured clinical metadata using Node.js, BullMQ, and MongoDB — designed for real-world scalability in modern healthtech.

Key Features

• Upload medical referral PDFs via REST — returns job ID instantly
• Background worker extracts clinical data asynchronously
• Redis + BullMQ for robust queueing, retries, and backoff
• MongoDB stores job state and results with 7-day TTL cleanup
• Bull Board for real-time job status, failures, and retries
• Extensible architecture — plug in LLMs, NLP, or ICD-10 mapping

Architecture Diagram

Architecture Pattern: Async Job Queue with Stateless API, Producer–Consumer Pattern, and Polling Client