Careful, hollowing out the undershank can trap residues from soap, cosmetics, or other chemicals the hand is exposed to. Along with the moisture involved it can cause long-term allergy reactions and sometimes a form of eczema. Best practice is not to do it, unless gallerywork in the side of the shank allows for venting. For balance/weight purposes, go ahead and hollow the ring but then solder in a sleeve. Easier than it sounds actually, just use the ring sizer to stretch a gold thin-wall tube up inside the ring for a swage fit before solder.
That being said, on the subject of milling hollow undershanks, it is possible but it is much tougher than you might imagine.
If you have the ring already cut, fixturing it properly on it's side can obviously be a problem as the rings are seldom even in width and/or cross-section. Then you have to either use something similar to a slot mill to go inside and mill the undercut, or use a rotary table at a 45 degree angle to use a ball mill.
If you try to mill the inside first, fixturing is easy, but you have to cut the ring to proper diameter inside, then mill the undercut using the above methods. The problem is then cutting the outside of the ring as you now have to have a mandrel to hold it on your rotary axis the same diameter as the ring size, and have it concentric, and have the wax indexed perfectly so the undercuts line up with the outside profile where you want it, and gripping it securely without breaking the now thin sidewalls of the wax. No one wants to turn and true the dozens of mandrel bushings necessary to do so, though I have on occasion made individual ones when necessary.
At the very least even if you did design the fixturing for doing so, it wouldn't always work as there are too many variables with designs and you wind up custom fixturing more often than not anyway.
I solve the problem by not undercutting at all, or simply doing so by hand in the finished wax as I then sleeve them so no one sees the sloppiness of the hollowing anyway. (It then has an "elite" touch not seen in "lesser" jewelry, meaning more implied quality and exclusivity and a higher price customers will gladly pay! I just do it 'cause it's actually easier, but I don't need to tell the folks writing the checks that!!!!) Make the sleeves from something ultra hypo-allergenic like 24k or .999 palladium, and people REALLY freak. It just so happens those two metals are really soft and malleable to swage in and to burnish the edges of, but I don't let that slip either. Lazy shortcuts that actually ADD to the rings price are so much damned fun!!!
The only exception to this is when I do very thin-walled Victorian filagree rings with open gallerywork requiring a hollow interior. Then I usually do the rings in two halves (usually left and right halves, done as two-sided flat cuts).
This brings up another trick; Two perfectly matched halves with fine details cannot be joined with normal wax pen work, as the seam shows. But if the two pieces have properly milled mating surfaces, a small amount of Cyanoacrylate (aka superglue) brushed on the mating surfaces with a small flux brush will bond together seamlessly and allows allows a few seconds of slip for perfect alignment. This will hold together for investing very well, and yes, it burns out perfectly cleanly. You can also use some hairsprays. Some of my rings have been assembled from as many as 8 separate wax pieces, allowing shapes and details no one could get from one solid block of wax, no matter haw many axes they had to play with! Look at real antique filagrees closely and you will see thats how they did it a century ago in the first place.
CA glues work for all sorts of fun shortcuts, like socketing round wax wire into properly sized blind holes drilled into your heads for uber-clean cast prongs, or mounting two halves of small double-sided pendants or earrings together, with none of the hassles of a two-sided cut.
Try envisioning the process of how these things have to be cut. It will make you a much stronger designer. It is too easy to compromise a design because you don't know the processes well enough to push the envelope of the possible. It is also too easy to simply design something that is uncuttable. At worst you will at least understand how to do the same thing in different ways and in doing so save lots of money, as it is also easy to ask for something that while not impossible, you will get charged a hell of a lot to have done when a different set of steps to the same result would be much shorter and less costly!
Try designing some multiple-piece assemblies, and use an STL assembler to view the completed unit virtually. Your life will get a lot easier and more profitable.