E-Madini

01 · Problem

Electronic waste piles up because there is no easy link between the people who produce it and the recyclers who can handle it. Producers do not know what their old devices are or where to send them, and recyclers have no simple way to receive pickup requests.

02 · Solution

E-Madini is an e-waste management platform, built as a PWA, that connects producers and recyclers. A YOLO11 model classifies electronic devices from a scan, with 92.6% test accuracy. The backend also handles pickup requests, location services with geocoding, and database migrations.

03 · Process

1. 01

   Research

   Looked at how e-waste moves from households and businesses to recyclers, and where that chain breaks.

2. 02

   Project requirements

   Defined what producers and recyclers each need, including device identification and pickup requests.

3. 03

   System design

   Designed the system architecture, the ERD and the data model for users, devices and pickups.

4. 04

   AI/ML pipeline

   I owned this part: trained a YOLO11 model to classify electronic devices, reaching 92.6% test accuracy.

5. 05

   API

   Integrated the model into the FastAPI backend through a scan endpoint, and built pickup requests, geocoding and database migrations.

6. 06

   User interface

   Built the app as a PWA so it works like a native app on any device.

7. 07

   Testing and deployment

   Tested the scan, pickup and location flows end to end, then released the app.

04 · Tech stack used

• YOLO11
• Computer vision
• Python
• FastAPI
• PWA
• Geocoding
• Database migrations

05 · Links