<p>In modern healthcare, early disease detection and personalized treatment play a crucial role in improving patient outcomes. Traditional diagnosis methods use techniques based on manual assessments and interpretations that introduce delays in diagnosis and in identifying a course of treatment. This paper proposes the Gated Recurrent Unit (GRU) neural network, enhanced by an Attention Mechanism, to model a sequenced progression of symptoms that could be leveraged to classify diseases and recommend medications. A GRU is a neural network designed specifically for sequential symptom data progressing through time. The Attention Mechanism simply contributes interpretability to the model, highlighting the most influential symptoms. The performance of the GRU with Attention in the benchmark assessment utilizing the dataset, resulted in accuracy of 99.19%. The GRU Attention model outperformed a BiLSTM, multiple transformer models (T5, GPT-3), as well as traditional methods of linear regression and/or machine learning methods. Additionally, the paper proposes a web-based diagnostic application that solves predictive disease and drug recommendations while also enhancing diagnostic utility, interpretability, and computational capability to acquire embedded patterns using data mining methods from large medical databases. The proposed system easily navigates the doctors so that they can input symptoms and receive instant feedback on current disease state assessment and possible treatment options, thus allowing them to market a clinical AI assistant on a global scale.</p>