BUILDER

PatientID+

Knows every returning patient by face — instant identification, instant records, zero re-registration.

Year2025
RoleSolo · Full-stack
StatusDXP Software · Production-ready

Context

Every hospital visit began the same way: re-registering the patient and re-entering their details by hand, even for people who had been there many times before — while each patient's treatment history stayed scattered and slow to pull up exactly when it was needed most. PatientID+ was built to end that repetition: recognise a returning patient on sight and surface their full record instantly. DXP handed me the brief during my first summer internship — the summer after my first year of engineering — as a deliberately challenging problem they trusted me to own from end to end.

Role & scope

Solo — sole developer, end-to-end across the full stack, during the DXP internship.

Problem

Manual, paper-based patient identification is slow and error-prone. Returning patients are registered from scratch on every visit, and their history is hard to retrieve at the exact moment a clinician needs it — friction that adds up across hundreds of visits a day.

Solution

A full-stack system that identifies patients in real time from a live camera feed using OpenCV facial recognition, then instantly retrieves the matching record — spanning the MongoDB data model and REST API through to the React front end. A returning patient is recognised on sight, and their full history is on screen before the conversation even starts.

Stack

  • React
  • Node.js
  • Express.js
  • MongoDB
  • Python
  • OpenCV
  • REST API

Outcome

Delivered and presented to DXP mentors, and integrated and tested against live security-camera feeds on site — the only intern-level project to ship as a standalone, production-ready system, recognised formally by mentors for its quality and completeness.

Links

  • Client project — Private (client work)