Automated grading has become a crucial tool in education technology, efficiently assessing large volumes of student work, offering consistent evaluations, and providing immediate feedback. However, current systems face challenges such as the need for large datasets, lack of personalized feedback, and a focus on benchmark performance over student experience. To address these issues, we propose a Zero-Shot Large Language Model (LLM)-Based Automated Assignment Grading (AAG) system. This framework uses prompt engineering to evaluate both computational and explanatory responses without additional training or fine-tuning. The system delivers tailored feedback, highlighting strengths and areas for improvement, thus enhancing learning outcomes. Our study shows the system’s effectiveness through evaluations, including survey responses from introductory higher education students, which indicate significant improvements in motivation, understanding, and preparedness compared to traditional grading methods. These results demonstrate the AAG system’s potential to transform educational assessment by focusing on learning experiences and providing scalable, high-quality feedback.