Nae Bother landrodger suprised one of your RPA's didn't spot it.
I can even suggest a method of sealing against it which might cut down your HE getting through seals so quick.
If you get your seal and put a water shroud over the top of it and fill it will distilled 100% H2O then freeze it down either using system temp or expanding gas over it. Then bring your system down to temp. The Ice will be denser than your seal so the He should take longer to get through it. And the interface where the cappilary action takes place will not have any air near it to form liquid O2. Vibration may be a problem breaking the seal between ice and surface but it could help out with some of the seals. If you can figure out a way of getting pressurised ice eg blue glacier ice with no O2 in it even better.
PS take all the glory for yourself, and do a patent search before you mention it to any one. Could be a nice boost to your pension. And in my child hood and to this day I don't think of MRI, it was and always will be NMR.