A group of students from the Faculty of Electrical and Electronics Engineering at Ho Chi Minh City University of Technology developed BKER-TEMPURA, a smart food frying system integrating computer vision, industrial robotic arms, and a Large Language Model (LLM). The idea came from surveys of restaurant chains, industrial kitchens, and food processing businesses, which revealed that the frying stage is a "bottleneck" in production lines. The surveys showed that dishes requiring advanced techniques, such as Tempura shrimp, pose challenges for industrial kitchens due to meticulous steps and the need for precision. Businesses must invest heavily in skilled labor, while consistent quality between batches is difficult to achieve. Manual sorting of materials on conveyor belts is prone to errors, and workers face occupational safety risks from hot oil and fumes. The system automatically identifies and classifies ingredients, performs frying processes with appropriate parameters for each food type, and allows users to monitor and control it using natural language. Computer vision identifies and classifies food before processing, setting corresponding temperature, time, and frying process to ensure uniform color, doneness, and flavor. Industrial robotic arms perform precise tasks, such as placing food into the fryer and coating with batter, replacing humans in complex tasks.