Modernizing DataRobot Models on Google Cloud Vertex AI

Executive Summary

A leading U.S.-based healthcare organization partnered with Factspan to modernize its machine learning ecosystem by migrating predictive models from DataRobot to Google Cloud Vertex AI. The engagement focused on mission-critical use cases such as patient payment forecasting, appointment cancellation prediction, and emergency department (ED) volume planning.

Factspan rebuilt 38 models using open-source algorithms like XGBoost, CatBoost, Prophet, ARIMA, and SARIMA, enabling greater transparency, scalability, and governance alignment. The migration achieved 100% success, with over half the models outperforming their original DataRobot benchmarks. The new Vertex AI framework reduced vendor dependency, improved explainability, and established a foundation for continuous innovation in healthcare AI.

Business Impact:

• 100% model migration success across 38 use cases
• 19 models outperformed previous DataRobot benchmarks
• 40% reduction in retraining and deployment time
• Full explainability and governance compliance via open-source transparency
• Significant cost savings by eliminating vendor dependency