GenderAI
From AI Experiment to Paid SaaS Product

The Starting Point:
Turning an AI Model into a Product

I learn best by building. I did not just want to understand how to train a model. I wanted to go the whole way: from dataset to a product someone pays for.

The task I set myself:

1. Source and clean a large dataset for training.
2. Train and run a custom neural network.
3. Build a full-stack application with a user interface and a public API.
4. Sell the service with a token-based payment system.

My Approach:
A Product, Not a Prototype

The Architecture

The backend runs on Node.js and Express and serves the API. The frontend is built with Angular.

From Model to Tool

Users get results right in the browser: enter a single name, or upload a whole CSV list and process it in one go.

Paying with Tokens

To turn the tool into a product, I built a token-based payment system with Stripe. Users buy as many tokens as they need; each processed name uses one. It bills the same way for the API and the web tool. And because nobody buys more than they need, the decision is easy. That helps conversion.

The Result:
An AI Product with Revenue

The result was a working SaaS application with paying customers and around 500 CHF in revenue. That proves not only that the technology works, but that someone pays for it.

What this project shows about how I work:

• I take an AI product from data to revenue.
• I build a full-stack application with Node.js, Express and Angular.
• I implement a payment system with Stripe.
• I connect data science to a product that sells.