Techniques for providing intelligent data profiling services are disclosed. A distributed computing system can include a data profiling service and a data recommender service. During an interactive data session, a user can select a data column of a data entity and request data profiling information for the data column. The data profiling service can compute the requested profile. The data recommender service can then determine a priority rank for the data columns of the data entity. Based on the priority rank, the data profiling service can compute a second data profile for a second data column of the data entity. The second data profile can be stored in a cache for rapid retrieval when the user requests the second data profile.