Billy Okeyo
Engineer / BuilderLive

Eyero

Offline cataract screening on a phone. An offline mobile app for early cataract detection in places where connectivity is a rumour.

Overview

Eyero is a mobile cataract detection app built for low-connectivity regions. The interesting constraint is the same as Tijafugo: the field is not a data centre.

The Problem

Early eye screening is hard to offer where clinics are far and the network is worse. A tool that needs the cloud at the point of use is not a tool.

The Approach

Put the model on the phone. Treat connectivity as optional. Keep the job narrow: flag a possible cataract early, then point someone toward care.

System / Product Design

The product is a screening aid, not a diagnosis. That boundary matters as much as the model.

Technology

Flutter · TensorFlow

Challenges

  • Offline is the product

    If inference needs a round trip, the app fails in the places it was built for.

Key Decisions

  • On-device TensorFlow

    The model travels with the phone. The network can stay down.

Outcome

Live at eyero.netlify.app.

What I Learned

  • In healthcare-adjacent tools, what the product refuses to claim is as important as what it detects.

Links