AI-based software pipeline transformation

A computer-implemented method comprising:

• receiving, at a conversion unit, a definition of a software pipeline in an entry point file and referenced templates, the definition of the software pipeline containing definitions of a plurality of tasks;
• dividing, by the conversion unit, the definition of the software pipeline into a pipeline definition skeleton and the plurality of tasks;
• determining, by the conversion unit, each of the plurality of tasks as allowed or disallowed according to a software pipeline output template:
• (1) when determining one of the tasks as allowed, a definition of the allowed task is unmodified, and
• (2) when determining one of the tasks as disallowed, applying, via an artificial intelligence (AI) unit, one of a plurality of functions to modify a definition of the disallowed task into one or more definitions of one or more converted tasks that are allowed according to the software pipeline output template and deliver identical functionalities as the definition of the disallowed task, the plurality of functions including at least one AI-driven function and at least one code-driven function;
• constructing, via a prompt construction unit, a first prompt by appending the definition of the allowed task and the one or more definitions of the one or more converted tasks to a first instruction string, the first instruction string including instructions to a generative model to re-arrange the definition of the allowed task and the one or more definitions of the one or more converted tasks into a new definition of the software pipeline;
• providing, via the prompt construction unit, as an input the first prompt to the generative model and receiving as an output the new definition of the software pipeline from the generative model;
• and
• delivering the new definition of the software pipeline to a user device.

A non-transitory computer readable medium on which are stored instructions that, when executed, cause a programmable device to perform functions of:

• receiving, at a conversion unit, a definition of a software pipeline in an entry point file and referenced templates, the definition of the software pipeline containing definitions of a plurality of tasks;
• dividing, by the conversion unit, the definition of the software pipeline into a pipeline definition skeleton and the plurality of tasks;
• determining, by the conversion unit, each of the plurality of tasks as allowed or disallowed according to a software pipeline output template:
• (1) when determining one of the tasks as allowed, a definition of the allowed task is unmodified, and
• (2) when determining one of the tasks as disallowed, applying, via an artificial intelligence (AI) unit, one of a plurality of functions to modify a definition of the disallowed task into one or more definitions of one or more converted tasks that are allowed according to the software pipeline output template and deliver identical functionalities as the definition of the disallowed task, the plurality of functions including at least one AI-driven function and at least one code-driven function;
• constructing, via a prompt construction unit, a first prompt by appending the definition of the allowed task and the one or more definitions of the one or more converted tasks to a first instruction string, the first instruction string including instructions to a generative model to re-arrange the definition of the allowed task and the one or more definitions of the one or more converted tasks into a new definition of the software pipeline;
• providing, via the prompt construction unit, as an input the first prompt to the generative model and receiving as an output the new definition of the software pipeline from the generative model;
• and
• delivering the new definition of the software pipeline to a user device.