This is a ground-up (well, almost, we used Numpy) implementation of a convolutional neural network (CNN) completed as the final group project of CS 491, "Neural Networks", at the University of New Mexico in Fall of 2024. Some of my contributions were backpropagation through the convolution (kernel) and max-pooling layers, tuning training hyperparameters, testing, and writing an accompanying proof of the sizes of the kernel and pooling gradients.