Fraud Guardian

Inspiration

I am inspired by the increasing number of fraud attempts through SMS, calls, and phishing messages that put everyday users at risk.
The goal was to create a simple yet powerful mobile tool that empowers people to quickly assess the risk of suspicious communication.

Fraud Guardian analyzes suspicious SMS, calls, or texts and provides users with an instant risk assessment.
The app highlights whether the message is likely safe or unsafe, explains the reasoning, and offers practical safety tips.

I have built Fraud Guardian as an Android mobile app using Kotlin, Jetpack Compose for the UI, and Ktor for API communication.
A backend model was developed to assess fraud risk based on text input, and the app integrates with it through REST APIs.
I also used Kotlinx Serialization for data transfer and Material 3 components for modern UI design.

One major challenge was setting up the Android project environment with Gradle, Kotlin plugins, and serialization dependencies, which caused several build issues.
I also faced difficulties syncing the app with the backend API in real time while ensuring performance and usability on the emulator.

I am proud that Fraud Guardian now successfully connects frontend and backend, runs on an Android device, and provides accurate, user-friendly fraud risk assessments.
Despite environment hurdles, I have built a working prototype under time pressure.

I have learned how to better manage Gradle and Kotlin build environments, align serialization across frontend and backend, and design a user-friendly interface with Compose.
I also deepened my understanding of risk detection models and how to integrate them into mobile apps.

Next, I plan to integrate real-time monitoring of SMS inboxes, multi-language support, and advanced fraud detection powered by machine learning.
I also aim to deploy the backend on cloud infrastructure for scalability and eventually publish Fraud Guardian to the Play Store.

Solo: Sweety Seelam

I contributed to the development of Fraud Guardian, a mobile application designed to detect and prevent fraudulent activities in real-time. My role involved building the fraud detection logic using machine learning, integrating secure authentication features, and ensuring a user-friendly interface for reporting and alerting users of suspicious transactions. I also worked on the app’s data pipeline, enabling reliable classification of potential fraud patterns and delivering actionable insights to end users.

Sweety Seelam

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.