An intelligent grading system that automates assignment evaluation using a multi-model AI approach. It features a robust OCR pipeline capable of reading handwritten submissions and uses Gemini 2.0 Flash (with Llama 3.3 fallback) to provide fair, unbiased scoring and detailed feedback. The project includes a professional-grade automated testing suite using Pytest, featuring unit tests and integration tests with mocking for external APIs (Gemini/Groq) to ensure reliability and CI readiness.

A hybrid console-GUI application that combines a robust CLI for betting logic with a Windows API-based graphical interface for realistic wheel spin animations. It features a comprehensive betting system (Straight up, Red/Black, Dozens) and uses Inter-Process Communication (IPC) via file I/O to synchronize game state between the logic and the physics-based rendering engine.

An AI-powered accessibility super-app developed in 24 hours at the Saarthi 2025 National Hackathon (Graphic Era Hill University). It features a smart AI Agent powered by Llama 3.1 (via Groq) that automates the entire user workflow through natural voice commands—handling navigation, accessibility adjustments, and job filtering. The platform also includes a multilingual voice assistant (Web Speech API) for real-time translation and an intelligent job matching system using TensorFlow.js to connect candidates with inclusive employers.

A modern facial recognition pipeline built with DeepFace and OpenCV. It supports face detection, alignment, verification, and attribute analysis (age, gender, emotion, race). The system integrates multiple state-of-the-art models like VGG-Face, FaceNet, and ArcFace, offering a hybrid framework for high-accuracy biometric authentication.