I doubt if it is possible to design something like this entirely from first principles with just a calculator and a clean sheet of paper, and no previous twincharging experience.
But a fair start might be to size the turbo compressors to provide sufficient airflow at about half the total estimated final boost pressure. These days exhaust turbines are fairly well matched for speed and energy to the compressor, with typically three alternative sizes of exhaust a/r housing being offered.
The larger size of a/r would most likely end up being most appropriate. The compressors will be run at high flow but at a relatively modest pressure ratio. The turbines will also need to run at high flow and a similar low pressure ratio.
The required turbos will be very large by any standard. A 400 CID engine with a roots blower is going to look to the turbos more like a 700-800 Inch engine. The turbos don't "know" the engine is supercharged, they just see flow. So both compressors and turbines are going to be enormous.
GT35's will be far too small. Something like GT42's with 1.34 a/r may be more like it, even they would be a fairly conservative choice. I know of one three liter (183 CID) twin charged engine in a road car that runs this particular turbo,(GT4288) and he reaches full boost at 3,500 Rpm.
GT35's sound about right for a flexible 400 inch straight turbo engine. Many guys fit GT35 1.0 a/r to this same three liter engine, and it works well. But things change very significantly when a supercharger is added.
The 8-71 has plenty of displacement per revolution, certainly more than enough to provide a very good range of efficient boost pressure adjustment for a 400 CID engine. But a suitable drive ratio for it will need to be selected experimentally.
Probably something that experience tells would "normally" produce around 4-5psi may be a good starting point. That will increase to something approaching double that, with the turbos feeding into it. So yes, I would expect a rather conservative underdrive ratio may be required to begin with.
Now how you wish to set all this up depends on the application and your preferences. The wastegates will limit boost to the 20psi set point, but that 20 psi will be made up of contributions from both supercharger and turbos.
Depending upon supercharger drive ratio, and turbine a/r the engine can behave quite differently with regard to boost profile, torque curve, and exhaust (turbine) back pressure.
Both supercharger and turbos could both be driven fairly hard to reach full boost below 1,500 Rpm, but that may not be what you want, as top end power would be compromised.
This is where it becomes tricky, it is fairly easy to change either the supercharger drive ratio, or the turbo exhaust housing a/r, or the wastegate boost set point. Not so easy to predict beforehand exactly what is going to happen with a particular untried combination.
There are three pressures that need to be monitored. Turbo compressor boost, combined total boost, and exhaust (turbine inlet) pressure. They will give an excellent idea of what is actually going on, and what needs to be changed to get to where you wish to go. It is all extremely flexible, and the engine can be readily matched to the application.
One handy instrument I often use is a dual boost gauge from a twin engined aircraft. It has two pointers that sweep the same scale. I have painted one pointer red, the other blue.
It is easy to monitor two pressures, and see the exact difference between two pressures as the vehicle is driven. The only disadvantage is that these aircraft gauges are fairly heavily damped, and the pointer movement perhaps not as rapid as it could be. But it is excellent for watching for example, where boost and exhaust pressures cross over on an ordinary turbo engine. Or for measuring turbo and total boost simultaneously in a twincharge setup.