Well, yeah. I stipulate all that you state. HOWEVER....
.... This is one guy, who claimed he worked on one software product, that could tip foreign elections.... and that this one form of software was now deployed here against us.
Your scenario is a graceful ballet of many, operating at their peak, with no problems.
He asserts that one software product does it all.
To my mind, that is pure nonsense.
If you couple that with the fact this guy was caught defrauding the government for some miracle Farsi translator vaporware, there’s utterly no there there.
Quite impossible to do as described, for all of the reasons that you mentioned. The "Whistle blower" is a fraud and a con artist.
Your scenario is a graceful ballet of many, operating at their peak, with no problems.
I have never actually seen a large software project go that well. I expect the scenario I described is extremely unlikely to have gotten past "blue sky" discussions. However, "Technical feasibility" would not be a reason for abandoning the idea.
Traditional vote fraud by "finding" trunk loads of ballots in a car is the way to bet on a risk assessment. Cheaper, more reliable, and does not required as much technical skill.