IMPLEMENTATION AND REAL-TIME PROBLEM SOLVING OF A DUAL AI CHAT INTERFACE USING LARGE LANGUAGE MODELS

The Dual AI Chat Interface is an intelligent platform that enables simultaneous, real-time interaction with two independently configured AI agents powered by Google’s Gemini 2.0 Flash Lite Large Language Model. The system leverages Transformer-based neural network architecture with persona-driven system instructions to deliver distinct conversational experiences — one structured and professional (Cognito), and one casual and accessible (Basic Bot). Unlike traditional single- agent chatbot implementations, this project incorporates a parallel asynchronous execution model using React Hooks for concurrent state management across two independent chat threads. Natural Language Processing (NLP) capabilities are handled natively by the LLM’s context window mechanism, which retains full conversation history to enable coherent multi-turn dialogue. The application further integrates local storage synchronization for session persistence, non-blocking API calls via the Google Generative AI SDK for responsive UI behavior, and spring-physics motion algorithms via Framer Motion for fluid user experience. This system provides a scalable, real-time, and personalized solution for dual-agent conversational AI, demonstrating how LLM-based architectures can be deployed in parallel without inter-thread interference.