As a micro-intern, I implemented a nearest-neighbor model to efficiently compare a user-input book cover with each image in the library to find the best match. The Macmillan catalog library consists of hundreds of thousands of covers as a set of images with ISBNs as their titles, which I handled using SciKit and ORB to preprocess images by generating embeddings for each after cropping dimensions. I researched different feature-detection algorithms and prototyped a model using Spotify's Voyager, an approximate nearest-neighbor search library for Python users. Afterwards, I utilized Voyager to efficiently find the nearest neighbor matches, generating a set of embeddings that are robust even with rotations, distortions, occlusions, etc. by researching different feature-detection algorithms.