Developed an end-to-end MNIST digit classification workflow for the KU-BIG/KUBIG_2025_FALL repository, focusing on a complete pipeline within Jupyter notebooks. Leveraging Python and deep learning techniques, the work encompassed data preprocessing, convolutional neural network model definition, training, evaluation, and result visualization. A file upload feature was integrated to streamline dataset and artifact management, supporting reproducible experiments and simplifying onboarding for new users. The notebook suite was structured to facilitate rapid prototyping and clear communication of results to stakeholders. This foundation enables scalable experimentation in computer vision and machine learning, supporting future extensions and iterative model development within the project.