Abstract: A data processing system implements receiving a definition of a software pipeline in an entry point file and referenced templates; dividing the definition of the software pipeline into a pipeline definition skeleton and tasks which definitions included in the definition of the software pipeline; determining each of the tasks as allowed or disallowed according to an output template, and applying a code-driven or AI-driven function to modify any disallowed task into converted/allowed task(s); constructing a first prompt by appending the allowed task and the converted/allowed task(s) to a first instruction string including instructions to a generative model to re-arrange the allowed task and the converted/allowed task(s) into a new definition of the software pipeline; providing the first prompt to the generative model and receiving the new definition of the software pipeline from the generative model; and delivering the new definition of the software pipeline to a user device.

Abstract: A data processing system implements receiving a definition of a software pipeline in an entry point file and referenced templates; dividing the definition of the software pipeline into a pipeline definition skeleton and tasks which definitions included in the definition of the software pipeline; determining each of the tasks as allowed or disallowed according to an output template, and applying a code-driven or AI-driven function to modify any disallowed task into converted/allowed task(s); constructing a first prompt by appending the allowed task and the converted/allowed task(s) to a first instruction string including instructions to a generative model to re-arrange the allowed task and the converted/allowed task(s) into a new definition of the software pipeline; providing the first prompt to the generative model and receiving the new definition of the software pipeline from the generative model; and delivering the new definition of the software pipeline to a user device.