Jump to a page or project

An experimental AI chatbot that simulates the teaching styles of Hitesh Choudhary and Piyush Garg using prompt engineering, without RAG, embeddings, or fine-tuning.
Persona AI is an experimental Generative AI application exploring how convincingly an LLM can simulate the communication and teaching style of a public educator using prompt engineering alone.
The application allows users to select between two AI personas inspired by Hitesh Choudhary and Piyush Garg and interact with them through a conversational interface. The system focuses on modeling characteristics such as vocabulary, tone, teaching methodology, technical depth, response structure, and Hindi-English communication patterns.
The core experiment deliberately avoids RAG, vector databases, embeddings, fine-tuning, and agent frameworks. Instead, each persona is represented through a carefully engineered system prompt created from analysis of publicly available content.
The application follows a simple request pipeline: the user selects a persona, the corresponding system prompt is loaded, conversation history and the new message are added to the request, and the request is sent to a pluggable LLM provider. The generated response is then rendered in the chat interface with Markdown and code-block support.
Persona prompts include communication rules, teaching behavior, vocabulary, language preferences, response patterns, and behavioral constraints. The system is also instructed not to fabricate personal opinions or experiences that cannot be supported by the public material used to construct the persona.
The project uses Next.js, React, TypeScript, and Tailwind CSS on the frontend, with Next.js API routes handling the backend integration. The AI layer is designed to work with different LLM providers such as OpenAI, Gemini, or Claude.
Key features include persona switching, session-based conversation history, Markdown rendering, syntax-highlighted code blocks, automatic conversation scrolling, and optional streaming responses.
Persona AI is intentionally designed as a proof of concept rather than a production persona engine. Its main purpose is to explore the capabilities and limitations of prompt engineering for AI persona simulation