The disclosed embodiments provide a system for processing data. During operation, the system obtains labels for entities found in portions of text in a first set of jobs. Next, the system inputs the portions of text and the labels as training data for a machine learning model. The system then applies the machine learning model to a second set of jobs to generate predictions of additional entities in additional portions of text in the second set of jobs. Finally, the system creates, based on the predictions, an index containing mappings of the additional entities to subsets of the second set of jobs in which the additional entities are found.