Pneumonia is a severe respiratory illness characterized by inflammation of the air sacs in one or both lungs, which can lead to serious consequences and increased mortality, particularly in vulnerable populations such as the elderly and immunocompromised persons. A prompt and precise detection is critical for the effective treatment and management of this illness. This study investigates pneumonia detection optimization utilizing sophisticated deep learning techniques, including Convolutional Neural Networks (CNNs), Automated Machine Learning (AutoML), and a hybrid CNN-Recurrent Neural Network (RNN) approach. Using chest X-ray images, we evaluate these models’ performance. The CNN achieves 96% accuracy, AutoML improves to 96.6%, and the hybrid CNN-RNN model reaches 97.72% accuracy. Results show the hybrid model’s superiority in complex medical image classification, highlighting the potential of these techniques to revolutionize pneumonia diagnosis and enable more efficient automated diagnostic systems.