Splitwise is a full-stack expense-splitting app built with Next.js and Express, featuring secure JWT-based authentication, PostgreSQL for data management, and a sleek UI with Tailwind CSS. This ongoing project includes comprehensive API documentation using Swagger for easy integration and scalability. Open for your contributons.

Library is a online ecommerce website with a big set of books in every genre. Built with love using HTML, CSS, JavaScript, React.

A 1:1 copy of your favourite video streaming platform. Netflix and chill. Built with love using JavaScript, React, Redux, Firebase.

Contributing to the FakeStake project, an open-source Next.js application, where I developed a Minesweeper game as a key feature. Additionally, I played a significant role in designing and implementing the user interface, leveraging Next.js's capabilities to create a responsive and dynamic user experience. My work involved optimizing component interactions and ensuring seamless integration with the app's core functionalities.

Led the way in building AutoBloc's user interface with React.js and Material-UI. Me and my team designed a smooth and user-friendly experience, adding extra security by integrating Metamask wallets. My focus on simplicity and the use of Material-UI components played a key role in making AutoBloc a success.

Developed an interactive weather forecasting application providing current and upcoming weather details for cities across India. Built using Next.js for a modern, server-rendered user interface and integrated OpenWeatherMap API for reliable weather data. Utilized React Query for efficient data fetching and caching. Implemented user-friendly features like real-time search suggestions for cities and dynamic UI updates based on weather conditions.

A frontend web developer for five months. Developed the web pages like landing page, member area, redirect page, TOS page and others with the fellow developers in the team.

Developed a ML model for driving and parking area detection using SegNet architecture, a type of convolutional neural network specifically designed for semantic segmentation, to accurately identify drivable and parkable areas in real-world images. Accuracy: 69.2% on a Cityscapes dataset.

Implemented LiDAR-based 3D object detection system for autonomous vehicles using YOLO v4 architecture and Velodyne HDL64 LiDAR(Light Detection and Ranging) sensor, achieving 67.1% accuracy for car detection, 68% accuracy for pedestrian detection, and 72.3% accuracy for cyclist detection.

Dive into the world of Computer Vision and Machine Learning. This application employs advanced techniques to interpret American Sign Language (ASL) through real-time hand sign recognition. Made custom dataset for each alphabet using openCV. Note that the video above is just for representation, my original video is missing out, will re-run the code and record the original output soon.

Designed a demo model of street lights which work on the input from Analog to Digital signal Converter(ADC). Using a ARM LPC2148 board and an ADC sensor to take the input, We could switch ON the light when sun light intensity is low and vice versa.

To take a break from my ongoing exams, I decided to put my Tailwind CSS skills to the test by quickly cloning my university's Learning Management System's login page. The results are showcased in the above video :)

In the above proteus simulation, we had used Universal Shift Resistors (USRs) for lifting up the digital ON(1) logic and also to bring down the digital ON(1) logic. The digital ON denotes the elevator position in the building and the LEDs denote the floors of the building. We had designed it for 12 floors.