a couple of comments :
1 - I am not sure this topic fits well in this type of forum, being more design oriented and less tool oriented. Do you know www.edaboard.com? There's plenty of great design advice there, if you want to try it.
2 - On a general note, being myself an industry designer, I know that there is a long path to go from a paper/thesis idea to a robust product for the market.
3 - Nevertheless, the path to go is *after* you have a basic piece in place : this basic piece is usually the result of implementation of theory available from books/publications.
4 - I have not clear what you mean with practical and repeatable instructions other than what I suggested. I am sure you can find more precise theoretical formulae for coupling and inductance with varying sizes/geometries, however I doubt you can find much more down-to-earth "recipes" than those. The rest of the path, as I said, is
a) spot what defines your RF transfomer, in terms of specification. Chapter 1 of the linked material plus what suits your needs.
b) building a first order structure from book formulae. (you should be able to do it after reading the first 2 chapters of the thesis I linked, and having done some extra research as needed)
c) simulating the structure with EM simulator (either 2.5 or 3D)
d) extracting a model. Chapter 2 and 3 of the thesis I linked.
e) using the model, gain understanding of which knobs act on your desired target performance. (RL,IL,BW,transform ratio,Lp,Ls)
f) N iterations between c) and e)
g) verify performance with interfacing blocks
5 - Transformers are *not* standard devices (standard == digital in these days), plus they require a substantial amount of know-how to be designed properly. No surprise you don't find them in the PDKs. I would warn you that even inductors, but actually everything in RF, are *not* "straightforward". Their use needs careful consideration each step on the way, skills, reviews. That's what we get paid for, however, so I guess it's fair.
I have worked only marginally with transfomers, basically building switching caps around a pre-defined and optimized inductive core, but I can maybe dig some extra info for you from my colleagues. You would then need to be specific however as far as which step of "the path" is posing you an issue.
Again, I am not sure this is the right place for this kind of discussion. I would be glad to continue, should the need arise, but I would like to get a feedback from the moderators whether this is O.K. or not.