I’m by no means an expert on Albanian, but I asked one of the authors of the section on Albanian in The Indo-European Language Family: A Phylogenetic Perspective about this yesterday, and he told me the following – though bear in mind that this was off the top of his head in an informal context, without looking at any sources at all, so take it with a grain of salt.
Splitting up gënjej into parts, -ej can be reasonably discarded: it’s roughly equivalent to Greek -αίνω (both from *-ani̯V-) and enjoys similar productivity in Albanian, so whatever the source of the verb, it’s very likely to have had -ej added on within Albanian.
The root, gënj-, likely represents genj- with loss of stress from the suffix. Albanian has rather a large number of loan words from Latin (meaning probably Balkan Romance a lot more often than actual Latin, particularly Classical Latin) where Lat. a appears as Alb. e, so the vocalism from gann- to genj- would not be surprising.
He was not certain offhand if there are any sure examples where a geminate /nː/ before non-front vowels developed into /ɲ/ in Albanian, but he couldn’t rule it out either. A geminate /nː/ before a front vowel or glide would certainly be more attractive and straightforward, though.
The only part of the development from ingannō to gënjej he thought seemed unlikely is that we would expect the prefix to show up and the Albanian form to be ngënjej, similar to ngushtë (← angustus) or ndjej (← indulgeō).
All in all, he said, a development from ganniō to gënjej would be quite unremarkable and likely in every aspect, while ingannō to gënjej is not implausible, but not without problems either.
(This does not take into account whether the simplex ganniō actually existed and was used at the relevant stage of Latin or Balkan Romance, but given that it seems to have direct reflections in several Romance languages, I’m guessing it likely did.)
Gjunjë, Gjymtyrë
which also come from Latin