The rise of hate speech on social media has major consequences for people, communities, and society. To combat this issue, we propose United Against Hate, a deep learning ensemble approach that leverages the strengths of Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) to detect hate speech and offensive language. Our method achieves state-of-the-art performance on a benchmark dataset, demonstrating significant improvements over existing work. Specifically, our approach yields a validation accuracy of 0.8802, F1-score of 0.8805, precision of 0.8816, and recall of 0.879, outperforming previous work by 1.4% in accuracy, 1.5% in F1-score, 1.6% in precision, and 1.4% in recall. By showcasing the potential of deep learning ensembles, our research contributes to the development of improved solutions for the detection and mitigation of the spread of hate speech over the internet.