RTI International has built a financial forecasting tool that we call Revenue Prediction Model (RPM). RPM ingests financial data from multiple databases, cleans the data, performs forecasting through cascading machine learning models, and generates daily reports in an interactive web dashboard that illustrates forecasted financial metrics for projects over time. RPM is used by financial analysts and business decision-makers at RTI to understand our financial health and outlook, which in turn allows us to make more informed decisions so that we can better serve our clients. In this presentation we will provide a system overview, describing our methods for creating a forecasting tool that generates daily reports and has automated quarterly model training, evaluation, and deployment to production. This talk will touch on Python in production, DevOps, MLOps, and process automation. We will describe the interaction between all of these pieces, our layers of modeling, and how we have set up continuous integration and continuous deployment (CI/CD) for pull requests and build and release pipelines and for automated model training, evaluation, and deployment to our development and production environments. This presentation focuses on system design and implementation.