While there are many different methods to performing predictive maintenance in aviation, most require sensor data. New aircraft do generate more in-flight data compared to older ones, but the data is still limited to a small percentage of parts. However, we do have a rich history of component removal data that we use for predictive maintenance. Using Azure Databricks to move on premise data to the cloud environment to leverage Spark (Azure Databricks) to efficiently create custom predictive models for hundreds of families of aircraft components. We use survival regression and business rules to approximate the risk of occurrence of a failure event, given the past occurrences of other events to each serialized component. At risk alerts are then generated and sent to maintenance planning to schedule the removal. With over 4,000 families of components across a fleet of over 400 aircraft, this type of predictive maintenance has already shown saving to the operation and the potential for much more. Predictive removals are tracked when sent to a shop for testing and repair. We currently have over 95% success rate with over $1.3 million in avoided operational impact costs in FY21.