Cyclin-dependent kinase 2 (CDK2) participates in eukaryotic cell-cycle progression. CDK2’s role in the cell cycle indicates that this enzyme is a target for anticancer drugs. Available high-resolution X-ray crystallographic structures and binding affinity data made it possible to construct machine learning models to predict affinity based on atomic coordinates. This work focuses on a workflow to integrate docking results generated with Molegro Virtual Docker (MVD) and experimental data to generate a neural-network model to calculate CDK2 inhibition. It integrates MVD and Molegro Data Modeller using Jupyter Notebooks and employs binding data from BindingDB ( https://www.bindingdb.org/ ) to train machine learning models. The model built using an integrated workflow shows superior predictive performance compared with a classical scoring function. This workflow can manage any target for which structural and binding data are known. The flexibility of the present approach can take experimental structures (e.g., crystallographic structures) and computational models (e.g., AlphaFold models). All CDK2 datasets and Jupyter Notebooks discussed in this work are available at GitHub: https://github.com/azevedolab/docking#readme .