Hope speech, in its mere definition, is a type that carries positivity towards views, empathy, and encouragement. It plays an important role in constructive communication and mental health. The paper discusses the detection and classification of hope speech using the HopeEDI dataset, which consists of various languages and contexts. Proposed machine learning models are Logistic Regression, Random Forest, and Support Vector Machines, SVM, combined with deep learning architectures such as CNNs with Fast-Text embeddings, RNNs enhanced with attention, and GRUs. Performing hyperparameter optimization is done through Grid-search in order to boost model performance. The performance is weighed using metrics such as accuracy, confusion matrices, and AUC-ROC score to assess the effectiveness of the models. It is hoped that this work will advance the cause of automatic identification of hope speech in furthering an empathetic and supportive digital environment.