Research on maintaining food grain sorting quality in the agriculture and food sector The Edge-AI pipeline is an automated grain sorting system leveraging computer vision and deep learning to classify rice grains based on quality. Traditional sorting methods are labor-intensive, inconsistent, and time-consuming, necessitating automation for enhanced efficiency. The model was trained on GrainSet, a dataset containing eight different rice grain classes, including normal, moldy, broken, and pest-infected grains. The goal of the project to implement light-weight models on other grain datasets to create a high throughput, less complex model to real time grain sorting. The reasoning behind is that to create a model which can run on edge devices like mobile phones or other IOT devices which will be cost efficient as well.