Designing a Persona-Driven Educational Chatbot

Sarada Prasad Malla

Master of Computer Applications
GIFT Autonomous College Bhubaneswar, Odisha, India, smalla2024@gift.edu.in

Allupati Ch. Patro

Assistant Professor

Master of Computer Applications
GIFT Autonomous College Bhubaneswar, Odisha, India, allupati@gift.edu.in

ABSTRACT:Programming education increasingly depends on interactive AI tools, yet most existing systems respond with formal, emotionally neutral language that reduces learner engagement over time. This paper proposes a persona-driven chatbot for programming assistance that addresses this gap by combining technical accuracy with a deliberately sarcastic conversational style powered by the Google Gemini API.The proposed system combines a context-aware conversational architecture with a stylized instructional persona that behaves as a sarcastic but highly skilled programming mentor. The chatbot is implemented using a decoupled client-server architecture consisting of a web-based frontend, an Express.js backend, and Gemini generative AI services. The system maintains contextual continuity through client-side chat history management while preserving API security through backend middleware.

This paper discusses the architectural design, prompt engineering methodology, context handling mechanisms, and user interaction workflow of the proposed chatbot. In addition, the paper outlines an evaluation strategy focused on user engagement, learning effectiveness, and conversational retention. The research demonstrates how persona-driven conversational AI can improve interaction quality and create a more engaging programming learning environment.

Keywords - Artificial Intelligence, Large Language Models, Gemini API, Educational Chatbot, Prompt Engineering, Human-Computer Interaction, Programming Assistance.