MediTrack — AI-Powered Clinic System

MediTrack is a complete patient management system built and deployed for a real Egyptian private clinic (El-Basma Clinic, Dr. Mohammed Basyony). Currently in active daily use, handling 200+ patient visits per day without performance degradation.

Engineered an Arabic voice dictation system using Groq Whisper + Llama. The doctor speaks in Arabic, and the system transcribes and auto-fills visit records with diagnosis, symptoms, and prescriptions.

Multi-layer transcription fallback ensures reliability: AI transcription → regex extraction → manual fallback.

Built a 14,000-entry fuzzy drug dictionary with per-doctor personal learning using RapidFuzz. The system recognizes drug names even with pronunciation variations, adapting to each doctor's speech patterns over time.

Deployed at a real private clinic, in active daily use

Handles 200+ patient visits/day without performance degradation

8.5/10 post-fix audit score across 20 files — rated production-ready by independent AI code auditor

Packaged for non-technical staff: one-click .bat setup, automatic backups, zero IT knowledge required

Django 5.2, Groq Whisper, Llama, RapidFuzz, SQLite (WAL mode)