TabeebAI

TabeebAI helps patients get to the right doctor. Instead of guessing which specialist fits their symptoms, patients describe what's wrong in plain language and the platform asks smart follow-up questions, flags anything that looks urgent, and recommends suitable doctors. It also keeps a patient's records and reports in one place, reading uploaded documents so the information doesn't have to be typed in by hand.

I built the whole thing end to end: the patient and doctor web app, the admin panel, the Node backend, and the Python AI service behind the assistant.

The problem

The project started from a real finding in interviews with doctors: patients regularly book the wrong specialist: a pediatrician for an adult, a generalist for something specialized. People can't reliably self-triage, so consultation time gets wasted, treatment gets delayed, and costs go up. On top of that, a patient's medical history is scattered across PDFs, lab reports, and images that nobody has time to read and reconcile.

So the goal was simple to state and hard to do well: understand what a patient is actually describing, keep them safe when it's an emergency, and route them to a doctor who fits, without pretending to be a diagnosis.

Architecture

In practice: a patient chats about their symptoms, uploads any reports they have, and the platform reads those reports, keeps the record organized, and points them to the right kind of doctor, with a video consult when they're ready.

Outcome

A working AI healthcare platform: patients self-describe symptoms and get matched to suitable doctors, urgent cases get flagged, records and reports are read and organized automatically, and consultations happen over video, all backed by an admin panel for approvals and oversight.