My initial advice here would be to proceed SLOWLY. The process of creating a modular system isn't simple, and you can't distill everything down to a few seemingly-apparent choices and slap something together off the top of your head. If you spend two weeks on MG and toss together a single build and say "that's that!"...well, that's what I would call a "hideous mistake".
There are two things you MUST know before proceeding down this long and expensive rabbithole:
1) Your own work. What about it suggests to you that you need a bespoke device to take it to the next level? IS this a proper course of action, or are you just following hype? Are there specific elements that a modular system would bring to what you do that you feel are necessary, given the expense and technical requirements? Those (and more!) are all questions you need to be asking yourself before proceeding here.
2) Synthesis. Do you know your synth architecture "basics"? As in, what HAS TO be there, as opposed to what might wind up being an expensive space-waster. Or moreso, what do you have to have to make a modular work on its most basic levels...while leaving stylistic considerations out of things at this point. That's important, because the best way to proceed here is to build a basic system, then start augmenting it. Plus, when doing this, OVERSIZE your rack selection...because the next step is to take the unworkable mess that results and pare it back down once more to arrive at something more workable. Remember: you can always shrink or enlarge your MG racks, but it's not so easy to do that if you've gone all in on hardware before getting sorted on here. And that "do/enhance/undo/repeat" method, done again and again, is how you whittle down 7000+ module choices (in Eurorack alone!) to a set that YOU can work with perfectly. But it's NOT an instant gratification exercise; expect to fiddle with MG builds for potentially months before the "real deal" pops up on your screen, the one that says "OK...whip out the plastic!".
So, it's not a simple process to create something like this, which is one reason that I and others here often dissuade starting users from bonking around in the Grid for a couple of days and then declaring that they've found "IT". Because, quite often, that "IT" is aptly-described by that term...as the result isn't what you'd exactly call a "synthesizer". Take your time, ask questions, explore, and again, take your time. It may have taken you (like many of us here) years or even decades to develop your craft; why should the process of creating an instrument that's ALSO "you" take a few days?