You've developed a custom ML model using scikit-learn for a project that processes large datasets. The training time has become a bottleneck, affecting the project timeline. To optimize the training process, you're considering migrating your model to Vertex AI Training. The project has constraints including a tight budget and the need for scalability to handle increasing data volumes. Given these constraints, what initial approach should you consider to effectively reduce the training time? (Choose one correct option)