Prashnarth: AI Driven Quiz Generation
Project Overview
Teachers, educational institutions, and corporate trainers waste hundreds of hours manually drafting test questions. Prashnarth automates this entire process by parsing source PDFs, DOCs, or raw text files, utilizing generative AI models to construct valid questions of varying difficulty levels, and hosting interactive live sessions.
Role & Execution
Prashnarth is a full-stack AI-driven assessment platform designed to automate quiz creation from structured and unstructured documents. It leverages Google Gemini API to generate diverse question types including MCQs, descriptive, and fill-in-the-blanks. The system supports real-time multiplayer sessions and adaptive learning modes, enhancing engagement and performance tracking. Built with a scalable backend and WebSocket integration, it ensures seamless real-time interactions. The platform significantly reduces manual effort in content creation while improving learning outcomes through intelligent automation.
Outcomes & Impact
- Reduced time spent on quiz creation by 90% for pilot educators.
- Supported 100+ concurrent students in live, real-time multiplayer lobbies.
- High accuracy in question generation with minimal hallucination rates.
Tech Stack Details
For secure, relational data management, quick migrations, and robust API logic.
Enables bi-directional communication required for multiplayer quiz lobbies.
Leverages cutting-edge LLMs for high-quality instruction following and contextual quiz drafting.
Key Features
- Automated document processing and semantic layout analysis.
- Integration with Google Gemini API for accurate context retrieval and dynamic MCQ, true/false, and short answer extraction.
- Real-time multiplayer lobbies using WebSockets for live group assessments.
- Interactive dashboards showcasing student performance analytics and common learning gaps.
- Adaptive difficulty scaling based on candidate response history.