This work aims to develop algorithms designed to detect stuttering in speech signals effectively. The background of this speech disorder is given with particular attention paid to its subcategories, i.e., extensions, blocks, sound repetitions, word repetitions, and interjections. A review of state-of-the-art machine learning-based approaches to detect and classify stuttering is presented. A deep model capable of detecting stuttering in the speech signal at the f1 measure level of 0.93 for the general class is built. Also, two baseline algorithms, i.e., Support vector machine (SVM) and k-nearest neighbors (kNN), are implemented to compare their efficiency with a residual neural network outcome. In addition, a series of experiments are conducted to study the impact of elements such as model architecture, data partitioning, amount of data, or use of individual features. The results obtained for the different categories of stuttering differ significantly due to the size of each subclass in the datasets used for the study and the annotation quality.