Normally, a DataPump "job" would be an instance of a DataPump "process" being executed, either manually, as a scheduled job, or via a script call to PumpAPI.
A "process" consists of a defined sequence of "projects". A "project" is an XML file created in Monarch that consists of three elements:
1. The path to the input file (in this case your text file)
2. The path to the model to use - there has to be one and only one model in a project.
3. The export paths for table and summary exports of the project.
So, thinking back to Lesson10 and the Betty's Music Store data, you might have a model "Lesson10.xmod", and an input file of "Classic.prn", and an export defined to export the summary "Sales by Media" as "Sales.xls".
You can tell DataPump to "monitor" a folder, so if you put the text file there, the process starts automatically. The model will still have to exist somewhere, however.
Hope this helps!