A home system prototype built with Proteus 8.9 and Raspberry Pi 4. This project showcases an IoT-driven home automation system that intelligently monitors motion, light intensity, temperature, and gas levels to enhance safety and energy efficiency. Simulated using Proteus and powered by Raspberry Pi, the system lays the groundwork for scalable, customizable smart home solutions—bridging the gap between industrial automation and accessible home tech.
This project implements an end-to-end image classification pipeline for detecting real and manipulated (deepfake- Deepfakes Face2Face, FaceSwap,NeuralTextures) images using the Swin Transformer (swin-tiny) architecture from Hugging Face Transformers and deployed using Gradio interface.
This work presents a novel framework that leverages Graph Neural Networks (GNNs)—specifically GCN and TGNN architectures—to predict U.S. economic recessions. The system integrates graph-theoretic modeling of macroeconomic indicators with deep learning, offering early-warning insights into structural economic shifts.
A passion project developed using the Godot Engine to explore game development fundamentals and real-time interactions. This 2D shooter features fluid controls, enemy wave generation, projectile mechanics, player systems, and responsive collision handling. The game draws inspiration from classic arcade shooters and was a hands-on exercise in learning Godot’s node-based architecture, GDscript and animation system.
More projects are on the way!