Given a dataset of images, how would you approach building a model to detect specific objects for Tesla Tecnologia e Comunicação?