receive a request to generate a job recommendation for a user, wherein the request comprises a payload of user data and an identifier of a type of recommendation to be made,

dynamically select a subset of machine learning models and a subset of input data sources for generating the type of recommendation from among a larger set of machine learning models and generate a communication path that interconnects the subset of input data sources with inputs of the subset of machine learning models and that also interconnects input and output of the subset of machine learning models in sequence with each other at a host platform to create a dynamically selected path between the subset of input data sources and the subset of machine learning models and within the larger set of machine learning models;

identify, via execution of the dynamically-selected subset of machine learning models, skill attributes of the user based on the payload of user data, wherein the hardware processor inputs a job description of the user from the payload of user data into the dynamically-selected subset of machine learning models which outputs the skill attributes;

determine, via execution of the dynamically-selected subset of machine learning models, a recommended job for the user, wherein the hardware processor inputs the skill attributes into the dynamically-selected subset of machine learning models which outputs an identifier of the recommended job;

dynamically select a template with predefined content embedded therein from among a plurality of templates based on the type of recommendation to be made; and

populate the selected template with a description of the recommended job and display, via a user interface, the populated template with the description of the recommended job.