The integration of artificial intelligence (AI) into drug discovery is transforming the field by significantly increasing efficiency, reducing costs, and improving the accuracy of predictive models. This chapter presents a comprehensive and reproducible AI-driven modeling pipeline, detailing key steps such as data preprocessing, molecular descriptor computation, and machine learning-based bioactivity prediction. The workflow leverages cheminformatics and computational chemistry techniques and employs advanced supervised learning algorithms—including random forests, gradient boosting machines, and neural networks—to analyze diverse molecular representations, such as SMILES strings, molecular fingerprints, and quantum descriptors. Special emphasis is placed on Quantitative Structure-Activity Relationship (QSAR) modeling, which establishes mathematical correlations between chemical structures and biological activities to facilitate lead optimization in drug development. In addition, the chapter addresses key challenges in AI-driven drug discovery, such as data quality, overfitting, and model generalizability, by implementing robust validation strategies, including scaffold-based and time-split validation. To enhance practical applicability, an interactive Google Colab notebook is provided, allowing researchers to experiment with AI-driven methods and apply them to real-world datasets. This framework provides a powerful and reproducible approach to accelerate drug discovery while ensuring transparency and reliability in predictive modeling.