I would think the T-stop difference from the F-stop would be a consistent offset, not different for each stop. So if the offset was 1/2 stop, it would be 1/2 stop for all aperture settings. But the only way to really know this is to have a lab do the T-stop testing and calibration for you. $$$
As was mentioned, I think the first suspect in your exposure difference would be the shutter.
Since you are talking about spending $$$ on this endevor...
First I would get all your camera shutters cleaned and adjusted to spec.
Then get the specs of the measurement of the final adjustment and variation for each speed.
Then you know how close it is to marked speed, for each speed.
Example final adjustment could be average of 1/3 stop slower than 1/125 sec (within spec), but with up to a +/- 1/3 stop variation for say 20 shots.
So, with the variation, the shutter could be 2/3 stop slow to dead on.
I think re-engraving would be difficult.
On the cheap, you could put a strip of white tape or label over the aperture ring and relabel it for the T-stops.
I think that is what I did back in the 70s when I was using a hand meter.
Today, with a computer, a color printer and some experimentation to work out the scale and spacing, a label would probably look pretty good.
As for the detents, I would have a tech remove the device making the detent and just use the aperture ring smooth, w/o detents. The detents would be in the wrong place anyway in relation to the recalibrated T-stops.