Self-sustaining deep learning software environment

A method comprising:

• feeding information about a plurality of software functions of a software process into a first machine learning model, the first machine learning model designed to identify flow among the software functions based on a simulation of the software process, the flow comprising a main trail corresponding to a first scenario;
• using a second machine learning model to generate a trail passage matrix, the trail passage matrix comprising all software functions of the software process and identifying the main trail as well as one or more alternative trails corresponding to the first scenario, through software functions of the software process;
• and
• passing the trail passage matrix to a deep learning neural network trained to identify an error occurring during execution time within the main trail of the software process and, based on the error, to redirect flow among the software functions to a particular alternative trail of the one or more alternative trails.

A non-transitory machine-readable medium storing instructions which, when executed by one or more processors, cause the one or more processors to perform operations comprising:

• feeding information about a plurality of software functions of a software process into a first machine learning model, the first machine learning model designed to identify flow among the software functions based on a simulation of the software process, the flow comprising a main trail corresponding to a first scenario;
• using a second machine learning model to generate a trail passage matrix, the trail passage matrix comprising all software functions of the software process and identifying the main trail as well as one or more alternative trails corresponding to the first scenario, through software functions of the software process;
• and
• passing the trail passage matrix to a deep learning neural network trained to identify an error occurring during execution time within the main trail of the software process and, based on the error, to redirect flow among the software functions to a particular alternative trail of the one or more alternative trails.