Dr. Docs - AI Powered Document Search and Retrieval (Team Project)

Made With:

An AI-powered document assistant leveraging FAISS, Sentence Transformers, and LLMs for semantic search. Users can upload documents in multiple formats and perform natural language queries for retrieval. Developed by a team of four and secured seventh place in Telangana at HackIndia 2025. Features include:

• Smart document search with summaries using semantic similarity.
• External resource retrieval: Converts natural queries to relevant links, articles, and topic suggestions from the internet.
• Google Drive integration: Enables natural language search for documents in Google Drive.
• Chat with Document: Interact with specific documents for contextual Q&A.

Haven - Your Virtual Safe Space (Team Project)

Made With:

Haven is a web-based mental health platform built to provide users with emotional support, community engagement, and AI-powered mental wellness tools. Developed by a team of four and awarded first place in the Web Development Domain at CBIT Sudhee Hackathon 2025.

• AI Mood Analysis: Detects user emotions and responds with empathy and suggestions.
• Penguin Assistant: A virtual pet that checks in and offers support throughout the day.
• Personalized Recommendations: Based on hobbies, interests, and preferences.
• Mood Capsules: Users can write letters to their future selves for motivation.
• Affirmation Mirror: Provides daily affirmations to build self-esteem.
• Gamification: Journaling streaks and stress-relief games built using Phaser.js.
• Mood-Based Chat: Connects users experiencing similar emotions using SendBird API.

Copy 'n Carry Google Chrome Extension - Smart Clipboard Manager

Made With:

Copy 'n Carry is a smart Chrome extension clipboard manager that captures and organizes copied content automatically. It groups copied text, links, and data by the originating browser tab, improving productivity and making retrieval seamless. The extension provides a persistent clipboard history, intuitive popup interface, and quick reuse features—all aimed at enhancing the user’s multitasking experience.

NEO Hazard Prediction Model

Model Made With:

Web Application Made With:

Using AutoAI from IBM Watson Studio, a machine learning model built on Extra Trees Classifier algorithm to predict if near-earth objects are hazardous towards Earth. The training is done on hazardous/non-hazardous classified objects from NASA's public database. The model is currently based on objects orbiting Earth, will further improve the model's accuracy to consider objects orbiting other planets. The model is deployed in IBM Watson Studio and integrated into a web application through a public API generated at the time of deployment. For real-time data prediction, the user can provide the ID of registered objects and through NASA's Open API the required input features are extracted, processed and fed into the model for prediction.

Voice Controlled MacOS Application Launcher

Made With:

A basic application launcher which opens any Mac OS application with the '.app' extension, Utilises the Mac OS 'say' command to convert text to speech and 'speech recognition' Python library to recognise user's voice with Google Speech Recognition API. To end the program user must say 'goodbye'.

Connect Four

Made With:

This Java program, developed using JavaFX, implements the Connect Four game. The program allows two players to take turns dropping discs into a 7x6 grid. The goal is to connect four discs vertically, horizontally, or diagonally. The application features a graphical interface where players can interact and see the game progress.