This project presents ReportEase, a Django-based web application designed to digitize and streamline crime reporting and case management processes. It enables users to submit FIRs online, track case progress through timelines, and securely manage evidence. The system includes an intelligent investigator recommendation feature using TF-IDF and cosine similarity to match cases with suitable officers. Additionally, real-time chat functionality using WebSockets enhances communication between citizens and investigators. Overall, the platform improves accessibility, transparency, and efficiency in the criminal justice system.

This project presents SkillLink, a web-based platform designed to bridge the gap between academic learning and practical work experience for students in Nepal. It connects students with internships, freelance projects, and entry-level jobs based on their skills and career goals. The system includes features such as secure authentication, role-based access, job posting, application management, and skill-based filtering for effective matching. By providing a streamlined and accessible platform, SkillLink helps students gain real-world experience while enabling companies to find suitable talent efficiently.

This project presents an intelligent news portal that combines fake news detection with personalized content delivery to improve reliability and user engagement. It uses machine learning models such as Logistic Regression, Decision Tree, Random Forest, and Gradient Boosting to analyze textual features and identify fake news in real time. Built using the MERN stack, the system also includes an admin panel for content management and analytics. Additionally, it recommends news based on user behavior and preferences, creating a more relevant and user-centric experience while promoting credible information.

This project presents a hybrid plagiarism detection system designed to identify both direct and paraphrased content. It combines traditional similarity measures like cosine similarity and Jaccard similarity with a supervised machine learning model that analyzes deeper linguistic features using NLP. While similarity metrics handle exact text matches, the ML model detects more complex forms of plagiarism. This approach improves accuracy and adaptability, making the system suitable for academic and content verification purposes.

This project presents Thrifter, a web-based e-commerce platform designed specifically for buying and selling second-hand products. Built using the MERN stack, it addresses the limitations of traditional e-commerce systems by focusing on sustainable shopping and personalized user experience. The platform includes features such as product browsing, cart management, and AI-powered recommendations to improve product discovery. By promoting reuse and eco-friendly consumption, Thrifter supports sustainable retail practices while providing an intuitive and user-friendly interface.

This project presents a personalized music recommendation system that suggests songs based on user preferences using Spotify data. It analyzes audio features such as danceability, energy, tempo, and valence to identify similar tracks. The system uses machine learning techniques like K-Means clustering and cosine similarity to generate recommendations, while integrating the Spotify API to fetch up-to-date song data. Built with a Next.js frontend and a modern backend, it provides an interactive interface for users to discover music, demonstrating the application of unsupervised learning and web technologies in recommendation systems.

This project presents a School Management System developed using Laravel and Vue.js to streamline academic and administrative operations. It includes features such as student and staff management, course handling, role-based authentication, and audit logging. With a modular and scalable architecture, the system ensures secure access, real-time updates, and efficient workflow automation. Designed using the Waterfall model, it enhances transparency, data integrity, and overall institutional efficiency.

This project presents a MERN stack-based job portal designed to improve personalized job recommendations and streamline recruitment. It incorporates role-based access for recruiters, job seekers, and administrators. The system uses NLP for resume parsing and cosine similarity for intelligent job matching, enhancing the accuracy of job suggestions beyond basic keyword filtering. Core features include job posting, searching, filtering, and personalized recommendations, making the platform efficient and user-friendly for modern hiring needs.

This project presents Shiksha Setu, an intelligent multi-source educational chatbot built using Retrieval-Augmented Generation (RAG). It allows users to upload documents or provide website URLs, which are processed, embedded, and stored using ChromaDB for semantic search. User queries are handled using a sentence transformer model, and relevant context is passed to the Llama-3.2 language model to generate accurate responses. Built with ReactJS, Tailwind CSS, and FastAPI, the system transforms static learning materials into interactive, queryable knowledge sources, enhancing personalized learning and academic support.

This project develops a microservices-based e-commerce platform using Node.js, React, MongoDB, and Docker, supported by a CI/CD pipeline. It addresses the limitations of monolithic architecture by improving scalability, maintainability, and modularity. The system follows Agile methodology and is tested using tools like Apache JMeter and Postman. By automating build, testing, and deployment processes, the CI/CD pipeline enhances developer productivity and system reliability. The study shows that microservices with CI/CD provide a more flexible and scalable solution for modern e-commerce applications.