And how exactly do you expect to find out if those plans have errors?
You can look at a cad drawing all day and not realize a hole is missing (been there) or that a hole is off by 1/4in (also been there) or designed themselves into a corner.
Designing a pcb isn't that difficult or expensive (especially non RBG ones), it's pretty simple wiring really, try hand wiring you figure it out quite fast. Transferring that to a pcb just means figuring out the software (same as CAD) and a hour or two of researching traces. There are not a ton of PCBs out there to just buy, most are designed for a specific case. This is one instance where if you want more than few it's almost always cheaper to design your own.
Any sort of milling operation is going to be expensive. They don't want to fire up the machine for anything less than a few hundred dollars due to setup fees and other costs involved, not to mention disrupting other jobs that pay more.
Deigning a case to fit a PCB is also backwards, sure you want to fit a pcb into the case, but that's relatively easy. The reason it's backwards is because pcb revisions are cheaper than case revisions and many measurements are known constants, such as witch positioning so really all you need to do is leave a bit of room around the perimeter and room for the controll and it's basically locked in how your PCB will fit. The other reason is time. You can get a pcb made in China and shipped in a matter of a couple days, it can be MONTHS before you get time on a CNC and again, this is per revision.
3d print designs are likely to be FAR, FAR more flawed than a design built for milling. Sure revisions are cheap, but most people working with 3d printers have pretty much ZERO engineering skills, I cringe at many of the 3d printer designs because of such poor design ideas being implemented or using whatever is on hand, which may not be on hand for you, you get things like "sorry for using 4 different screws to attach ___, it's what I had". There is also the fact that something designed for 3D printing may not even be possible on a milling machine or vice versa. You can design for both or even fudge a little but it's a different design philosophy (as well as tolerances) for each.
Having said all that you should also know CAD is actually the easiest part of designing a keyboard or pretty much anything really, it's cheap/free, you can learn at your own pace and you can do it yourself. I'm not saying people who do cad are underpaid or even that it's easy, I'm saying the grand scheme of getting something into your hands it barely qualifies as step 1 of 100. It doesn't matter how simple you think it is, complex isn't really that much more than cheap once it rolls out of CAD, lack of experience is what costs money.
It sounds like you have no design or manufacturing experience, if that's the case I urge you NOT to go down this path. You will end up with a horrible keyboard that you spent $500 on, or you will have a nice keyboard worth the price of a decent used car, that's not an exaggeration. If you are truly determined though, buy a 3d printer, hand wire a custom board you design, THEN consider the next step. Believe it or not it will actually save a lot of time and money.