Skip to content
Wincel Crusit
Projects

AVERA

An Android app that checks whether a handwritten signature is genuine, and shows which parts of the signature drove each result.

00:00:00:00
My role
Project lead, ML and frontend
Team
[team size]
Timeline
Apr to Nov 2026
Built with
React Native, PyTorch, TensorFlow

The problem

[Who checks signatures today, how, and what goes wrong]

What we built

A Siamese neural network compares a signature against a reference. Explainable AI highlights the regions that decided the result, so the person checking can see why, not just a yes or no. [Add the web side]

How I led it

[How you split design, backend and ML work, and one decision you made as lead]

Results

[Accuracy, defense outcome, or user feedback]

Phone screen
Phone screen
Phone screen
Web screen or model visual