Moaz Reyad developed a convolutional neural network model for hematologic disease analysis within the apache/singa repository, focusing on delivering a complete, extensible architecture for medical image classification. Using Python and leveraging deep learning and neural network expertise, Moaz implemented convolutional, batch normalization, pooling, and fully connected layers, integrating a softmax cross-entropy loss function to support robust training. The work included a factory function for model instantiation and defined a set of supported optimizers, streamlining experimentation and standardizing training workflows. This contribution provided a reusable component that accelerates rapid prototyping and supports future production evaluation for hematology image analysis in Singa.