Smart AI Payment Tracker

Overview

This is a full-stack project with a machine learning model. Our team created this project in a hackathon. We were tasked with creating a smart AI payment tracker that sends notifications to customers before the due date if they have delayed payments in the past. When a customer makes a transaction, the data is processed by the machine learning model to predict if the payment will be delayed, prompting an automated notification to ensure timely payment.

See Project:https://smart-payment-tracker-frontend.vercel.app

Model Training

I explored training a model in TensorFlow.js. By querying ChatGPT for guidance on the training process, I built a Feed-Forward Neural Network (FNN/MLP). The architecture features four layers: an input layer, two hidden layers, and an output layer. The first dense layer takes 2 inputs with 10 neurons and uses the ReLU activation function to learn patterns and relationships. The second dense layer extracts features using ReLU with 5 neurons. The final output layer uses the Sigmoid function for binary classification, outputting 0 for on-time payment or 1 for a predicted delay.

See Code:https://github.com/Joshijay1917/payment-prediction-model

Problem & Solution

In this system, there are two core issues faced by small-to-medium enterprises:

Problems:

For SMEs, it is difficult to manage and track customer payments manually.
It is time-consuming to send reminder messages to each customer individually through WhatsApp, SMS, or Email.

Solution:

A modern, clean, and intuitive dashboard makes it easy to track and manage customer payments.
A machine learning model identifies customers likely to delay payments, and APIs automate the notification process.

I am interested in backend development and continuously learning about it. If you find a backend developer then feel free to ask.

Quick Links

Home Skills Experience Projects

Contact On

+91 9429248465

jayjoshi1912007@gmail.com

@jayjoshi19