DataRobot does not have the traditional survival models like Cox Proportional Hazards model. But, you can convert the problem as a binary classification problem, which can be done using DataRobot. This can be done by breaking down the "patient" level data into "patient-time" format by creating multiple rows for the patient at discrete time intervals. The target for the binary classification model would be if the patient survives a particular time interval. For uncensored patients, the target would be 0 for all the time intervals except the last time interval while for censored patients, the target would be 0 for all the time intervals. The predictions from the model would indicate if the probability of the event happening a particular time interval which would be instantaneous hazard probability. Instantaneous survival probability would be complement of instantaneous hazard probability and you can recreate survival curves by cumulating the instantaneous survival probability.