Abstract: A computer hardware system including an observer and an artificial intelligence (AI) agent is configured to perform the following. Event data corresponding to a particular predefined observed task having a plurality of discrete steps are collected within an observation matrix by the observer. The event data is structured by the observer into a plurality of task data clusters. The task data clusters are encoded by the observer with respective keys. The task data clusters with the respective keys are stored by the observer within a collection database. A task documentation file including textual descriptions respectively corresponding to the plurality of discrete steps of the predefined observed task is generated by a large language model within the AI agent and using the respective keys. The event data within a particular task data cluster corresponds to a discrete step of the particular predefined observed task.

Abstract: Mechanisms are provided for organizing food items within a refrigerator system. The mechanisms detect the presence of a food item within a refrigerator system. The mechanisms retrieve a food item profile for the food item and determine, based on application of one or more rulesets to parameters of the food item specified in the food item profile, a placement for the food item at a goal location within a compartment of the refrigerator system. The goal location comprises a specification of a shelf in the refrigerator compartment and a location on the shelf. The mechanisms control mechanical movement mechanisms of the refrigerator system to move the food item from a current position within the refrigerator system to the goal location.