Campus dining systems often rely on manual processes that lead to long queues, inefficient service, and significant food waste. This paper presents Qless, an AI-driven platform designed to improve efficiency and user experience in campus canteens through intelligent automation. The system integrates natural language processing for conversational ordering, machine learning-based demand forecasting, and an adaptive token-based queue management mechanism. To evaluate the system, a synthetic dataset simulating real-world canteen operations over a three-month period was generated. Experimental results show that Qless significantly reduces average waiting time by approximately 65% and improves service throughput by 38% compared to traditional FIFO-based systems. The forecasting model, based on LSTM networks, demonstrated strong predictive accuracy, enabling better preparation planning and reduced resource wastage. Built using a microservices architecture with modern web technologies, Qless is designed to be scalable and easily deployable in institutional environments. Beyond operational efficiency, the system also contributes to sustainability by minimizing food waste through predictive planning. This work highlights how integrating machine learning with real-time system design can transform conventional service environments into intelligent, adaptive systems.
Vaidya et al. (Sun,) studied this question.