PrepMind AI: A RAG Based Intelligent Learning Assistant for Competitive Exam Preparation

Department of Information Technology, Pune Vidyarthi Griha’s College of Engineering & Shrikrushna S. Dhamankar Institute of Management, Nashik, India

^* Corresponding author: This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

Competitive Examination such as UPSC and MPSC involves rapid and correct responses, yet majority of trainees rely on conventional learning or coaching schools which are more expensive. Learners with access to the current digital tools will risk the failure when answering handwritten questions, text recognition and proper adjusting to their individual learning requirements making the difference between accessibility and quality answers a significant one. In order to address these weaknesses, this paper presents PrepMind AI, and AI powered learning system that is aimed at offering an immediate and accurate response using an integrated pipeline of Optical Character Recognition (OCR), Retrieval Augmented Generation (RAG), and state-of-the-art Large Language Models (LLMs). This research aims at producing a platform that is able to respond to text and image queries, finding the correct answers to such queries in the verified sites, and generating the correct answers to the particular students. In essence, it works by OCR to extract text in the form of pictures and a tool named RAG to extract the correct information and images in PDFs. The AI then responds to you in the same way a real human being would respond. The entire idea behind PrepMind AI is that it will be much easier to study for exams, and accessible to all students, regardless of their origin.