[PeterisP]

IMHO those jobs are exceedingly rare, because all the roles with difficult algorithmic problems seem to have the grind around them as the majority of work that needs to be done. You mention data science, well, if you're doing serious data analytics two thirds or more of your time is converting data from one form to another and handling its flow; if you're doing cutting edge AI research, most of your time is devoted to 'plumbing' required to run these experiments, and only a minority to the algorithms you're researching. The same is pretty much everywhere, so if you find a job where that does not apply, that's quite exceptional.