PU Buddy: An AI-Powered Education Career Guidance Chatbot for Parul University

Authors:

Rudra Sandeep Patel, Prof. Janmejay Vishwakumar, Mr. Jigar Pathak

Abstract— In the era of digital transformation in education, providing real-time, personalized guidance to prospective students remains a significant challenge for large universities. This paper presents PU Buddy, an AI-powered Education Career Guidance Chatbot designed specifically for Parul University, Vadodara. The system leverages the Google Gemini 2.0 Flash large language model (LLM) through the OpenRouter API to provide accurate, context-aware responses about courses, fees, hostel facilities, scholarships, admissions, and campus life. The chatbot is built usinga Flask (Python) backend and a modern HTML/CSS/JavaScript frontend, incorporating advanced features such as multimodal image analysis, voice-to-text input via the Web Speech API, session-based conversation management, and structured HTML table outputs for data-rich queries. A comprehensive knowledge base derived from the official Parul University 2025–26 Brochure is embedded as a system prompt to ensure domain-restricted, hallucination-free responses. Evaluation results indicate that PU Buddy successfully reduces student inquiry response times by an estimated 85%, handles parallel sessions with contextual continuity, and gracefully degrades to a fallback offline guide when API limits are exceeded. The system also includes a pixel-perfect university website clone with 570+ pages, serving as a standalone digital information platform. This pap.

Keywords - AI Chatbot, Career Guidance, Large Language Models (LLM), Natural Language Processing (NLP), Flask, Gemini 2.0, OpenRouter API, Education Technology (EdTech), Parul University, Multimodal AI