During December 2024, Zilong Xie developed a convolutional neural network model for image processing within the apache/singa repository, targeting the TED CT workflow. Leveraging deep learning and neural network expertise, Xie implemented the model in Python, incorporating convolutional, pooling, and fully connected layers with softmax cross-entropy loss. The work included configurable training and distributed options, allowing flexible experimentation and scalable analytics. Utility functions were added to streamline model instantiation, supporting reproducible research and rapid prototyping. Although the contribution spanned a single feature over one month, the implementation demonstrated depth in model design and extensibility for future image analytics tasks.