I can't speak for NYC, but the salary noted for SF (just under $120kpa) is almost certainly base pay for the most junior engineers you can find who aren't interns. Three years ago my team hired an intern into a FT position for $100kpa base, and it was widely acknowledged that he was grossly underpaid, because he was far below the target midpoint for the entry-level pay band (that company was targeting the 65th percentile, and he was below it by a double-digit percentage).
Companies big enough to need real HR departments typically prescribe "pay bands" for every job position, e.g., a "Software Engineer II" might have a band between $100kpa and $130kpa, and to go above that band an employee either has to be promoted or receive a special exemption. Typically the midpoint of these bands (e.g., $115kpa in my example) are targeted to some percentile of one or more salary surveys done by companies like Aon. Those surveys are what set the "market rate", generally. At least for non-FAANG companies, though I imagine FAANG companies use them as well, but with (much) higher multiples.
Yeah, from what folks are saying here it seems I made a mistake using Glassdoor for the analysis. In the future, if I do something like this again, I'll probably use Levels.fyi or something else.