I like simple. The more complex you make things the harder it is to engineer everything to work correctly.
Unless you have a skid steer to start from scratch with, you are buying a lot of $$$ parts. You specified some guidelines but I'm going to make some arguments against the platform design you want to use.
Hydrostatic drive is proven to work but they have some major drawbacks. For one, if you run out of oil, they sit until you get more. They take warm up time before you can drive in the cold. They require cooling, since they are making heat all the time when under load. When the oil is cold, you have to be careful, when the oil gets too hot, it burns up or causes failures; water contamination from an unforseen dunking could be $$$ burning up motors, pumps and corroding everything in sight unless you can get all the water out.
Automatic transmissions are also pretty iffy for me. While they are convenient, they waste HP and make heat. Even if you got a lockup torque converter, there will be inefficiencies and more moving parts. Although the modern ones are pretty reliable and strong, they are a very inefficient way of turning fuel into heat and when something fails, they may strand you right there.
Turning brakes that are dry will not work correctly without some very big components to dissipate heat. Wet brakes are strictly tractor type stuff and bring huge weight penalties with them.
Variators are a cool design but the modern versions are big $$$$.
Personally, I favor the Tucker 4 track design for anything I built at home. Simpler, can be made out of junkyard parts and pretty robust.
Although I am working on a Sno-Trac, if I built one at home, I would build a Tucker clone.