If the film is completely transparent: no numbers on the edges and, critically, no fogged leader, then you've never seen the developer. But has seen fixative, or it would look like an undeveloped film (possibly with the anti-halo layer removed), which is a bit gray (and will darken over time as the light hits it).
As Alan Marcus says, if the leader is fogged and there is border printing, then he has seen the developer and there is a problem with the camera, or he has never been to a camera.
It is difficult to see how the completely transparent housing could occur, especially for the bottom roll in a tank. If you are waving for investment, then really hypocrisy I think some chemistry will happen in every movie, even if not evenly. If you are stirring otherwise, then it is possible that upper part roll in the tank to be partially or completely out of the developer, but even then it is difficult to see how it will not splatter when shaking, and when filling and emptying the tank.
If the movie has seen any development (and was exposed), then there would be Some image there: it can be very pale, and it can be partial if it is by splash, but it will be nothing. Their reported dilutions and times (the first) look good for D76 for common emulsions: obviously, things vary to get the best negs, but this is an OK ball park: there would be a reasonable image.
And also, we know what the repairman saw: how can that work?
As described, I find it very difficult to see how this could have happened, unless the film has never been coated with emulsion or something like that: that should be detectable by watching the film: if there is no emulsion, both sides will be bright. Is possible to wash the film emulsion, but you need Really hot water to do that, and even then there would be fragments.
It would be useful if you provided details of the tank, agitation, etc.