"It seems to me extremely implausible that a parameter could exist which could cause software development to be 97% male in some countries but >50% female in some others (which is what we see). This is surely cultural."
After the experiences of the 20th Century, most people don't dare approach 'scientific' racism as an explanation for the difference between cultural behaviours, but some will still happily have a go at 'scientific' sexism. But the same reasons that we should discredit the former - before any moral objections, just on the basis of complete scientific unsoundness - also apply to why we should reject the latter.
We should be extremely sceptical towards claims founded on what the brain (either in its biological 'design' or neurally-manifest 'behaviour') can 'say' about how we act. The best we can demonstrate is that certain emotions or abilities have a _correlation_ with activity in certain parts of the brain - and _why_ this should even be the case, and how that works, we can only hypothesise and speculate; on the basis of empirical psychological studies, by studying (with caution) how damage to certain parts of the brain has affected behaviour, and by using neural imagery (which is still in its infancy).
This is good enough to help in diagnosing and hopefully developing treatment for brain-related illness or damage (and holds promise when used in this context), but I don't think it at all licenses us to make the exaggerated claims commonly made by pop neuroscience, which are too often scarily similar to those made by Victorian phrenologists (with as little genuinely scientific justification). We simply do not know how the brain works.
Neurodeterminism is far cry from 'explaining' how the brain works, much less the _mind_ which is more than the sum of the neural activities of the brain at any one snapshot in time - or even the sum of the neural activities, were we able to accurately record them in sufficiently high resolution (which we can't), taken across a _period_ of time.