• GPU-based Image Pipeline: Implemented real-time image preprocessing and postprocessing using WebGL shaders to offload computation from the CPU and improve responsiveness. • Local Feature Detector/Descriptor in WebGL: Built custom implementations of local keypoint detectors and descriptors (e.g., FAST, ORB) using fragment shaders for fully GPU-side processing. • WebAssembly Modules: Ported and optimized traditional C++ detector/descriptor algorithms to WebAssembly for high-performance execution within web workers, enabling parallel processing and thread-safe execution. • Modular JavaScript Architecture: Utilized web workers for asynchronous processing and ensured scalability across devices.

Gesture-controlled fruit slicing game for mini ninjas!

A smart resume-to-job matcher that helps applicants beat the ATS and land interviews.

An AI-powered web app that helps users generate professional-looking logos based on input keywords and style preferences. Uses a pretrained style-transfer model with a lightweight frontend for rapid iteration and preview.

Machine learning-based Android malware detector using Natural Language Processing techniques. This project extracts and processes app permissions and API call sequences as features, then applies ML models (e.g., SVM, Random Forest, etc.) for classification. Focuses on explainability, lightweight detection, and robustness across app categories.

Dailio is a personal productivity and self-improvement app designed to help users organize their day, build better habits, and track progress—one day at a time. It combines essential daily tools like alarms, timers, to-do lists, notes, and daily challenges into a simple, motivating interface that keeps you focused and accountable.

Project-Based Learning hub