P

python

Fully offline document Q&A system using PyMuPDF4LLM, Ollama, and llama.cpp. Dual LLM backend with pre-commit quality gates.

RideSync is a full-stack ride-sharing web application that connects drivers and passengers traveling in the same direction, helping users reduce travel costs, save time, and promote sustainable transportation.

MedMinder is a smart medicine reminder application designed to help users take their medications on time and maintain a healthy routine. Many people, especially elderly patients and individuals with busy schedules, often forget to take medicines regularly, which can lead to health complications. MedMinder provides an easy and accessible solution to this problem through reminder notifications, voice-guided alerts, and medication tracking.

The application allows users to add medicine schedules by entering medicine names, dosage details, and reminder times. At the scheduled time, the app sends browser notifications along with voice-guided instructions to remind users to take their medication. The interface is designed with accessibility in mind, featuring large high-contrast buttons and a simple layout to support elderly and visually impaired users.

MedMinder also includes a medication history tracker that records whether medicines were taken or missed. This history can be shared with caregivers or family members to help monitor the user’s medication routine. The project focuses on improving healthcare management using simple web technologies while providing a user-friendly experience.

The application is developed using HTML, CSS, and JavaScript, along with browser APIs such as Notifications API and SpeechSynthesis API for reminders and voice assistance.

RAG Chatbot built using LangChain, FAISS, Groq LLM, and Flask. Supports PDF document ingestion and intelligent question answering through Retrieval-Augmented Generation.

AI-powered Smart Hospital Complaint Management System built with Streamlit, Pandas, and NLP.

DeepFake Shield is an AI-powered platform designed to detect and analyze manipulated digital media, including images, videos, and audio. The system uses machine learning and forensic analysis techniques to identify deepfakes, synthetic content, and media tampering. It provides users with trust scores, authenticity reports, real-time alerts, and both online and offline analysis capabilities, helping individuals, organizations, and law enforcement combat misinformation and digital fraud.

AI-powered PDF Chatbot using Retrieval Augmented Generation (RAG), FAISS, and OpenAI APIs.

Civic pulse is an Ai-powered Civictech platform that helps authorities identify , monitor , and prioritise civic issues using data-driven insights and analytics

DisasterGuardAI is an AI-powered disaster reporting and rescue coordination platform built with Streamlit.

Attendance Tracker is a digital solution designed to automate and simplify attendance management for organizations, educational institutions, and teams. The system enables users to record, monitor, and analyze attendance data efficiently while reducing manual effort and errors associated with traditional attendance registers.

The application provides features such as employee/student check-in and check-out, attendance history, leave management, real-time attendance monitoring, and report generation. Users can mark attendance through various methods, including manual entry, QR code scanning, GPS location verification, or facial recognition, depending on the organization's requirements.

Administrators and managers can access dashboards that display attendance statistics, late arrivals, absences, and workforce availability in real time. The system also supports automated notifications, attendance correction requests, and integration with payroll or academic management systems.

Advanced versions of the Attendance Tracker may include AI-powered analytics, geofencing, anomaly detection, and predictive insights to help organizations improve attendance compliance and operational efficiency.

RescueBite AI

RescueBite AI is an intelligent food rescue and redistribution platform designed to reduce food waste and combat hunger. The platform connects food donors such as restaurants, bakeries, hotels, supermarkets, and event organizers with NGOs, shelters, and volunteers through a real-time donation network.

Using AI-powered food classification, freshness assessment, and smart matching algorithms, RescueBite AI ensures that surplus food reaches people in need before it expires. The system streamlines the entire donation lifecycle, from donation creation and verification to pickup, delivery, and impact tracking.

Key Features Food donation creation and management AI-powered food classification Food freshness and urgency assessment Smart NGO matching Volunteer assignment and coordination QR-based pickup and delivery verification Real-time notifications Donation history tracking Impact analytics dashboard Food waste reduction insights Mission

To build a scalable and transparent food rescue ecosystem that transforms surplus food into meaningful meals while reducing environmental waste and supporting communities in need.

Team

Team Name: Vandal Institution: ICFAI

Vision

A future where no edible food goes to waste and no community struggles with food insecurity due to inefficient distribution systems.

DocSage — Streamlit RAG chatbot for document Q&A with PDF structured-qa and Gemini RAG for other sources.

Local-first RAG chatbot for university documents using PyMuPDF4LLM, FAISS, Groq and Llama.cpp.

Local PDF RAG system using TinyLlama, FAISS, Streamlit, and Sentence Transformers.

Local-First PDF Q&A Assistant built using PyMuPDF4LLM and Llama.cpp. Supports PDF upload, document retrieval, and question answering using a local language model.

Production-ready RAG chatbot built with Streamlit, LangChain, FAISS, and local LLM support. Features document Q&A, PDF processing, and AGPLv3 compliance.

Local PDF Question Answering System using TinyLlama, Llama.cpp, Streamlit, and Retrieval-Augmented Generation (RAG).

Smart Water Leakage Detection System built with Python and Streamlit for monitoring, reporting, and analyzing water leakage incidents.

Multimodal Retrieval-Augmented Generation (RAG) application built with Streamlit, LangChain and ChromaDB for intelligent document question answering.

A dual-mode RAG (Retrieval-Augmented Generation) chatbot. Mode 1 uses hybrid search (pgvector + BM25) with Ollama for fully local, free LLM inference. Mode 2 uses a lightweight Find-Retrieve-Answer workflow with PyMuPDF4LLM and llama.cpp for structured PDF Q&A with zero embeddings or vector database.