Build a job embedding space

1) build a job embedding space in what ever way you identify as best based on some distance metric of your choice 2) use the embedding to randomise jobs based on a specified blur distant metric The model should function such that; 1) one specifies a blur distant metric 2) one specifies an input job 3) the model takes the blur distance and the job and uses the embedding to output a random job selected from the local neighbourhood of the job specified where the neighbourhood the random job is selected from is defined by the specified blur distance metric When referring to the local neighbourhood we are referring to the local hyperspace of the model where there are similarly-named or themed jobs, such as "gardener" and "landscaper" for example