I think that the best way to make progress on those ideas is to make small proof-of-concept versions of them first.  I did not know how to make a platformer, or an FPS, or an RPG, so I made little ones.  The RPG plus the other experience I had gained by that point was enough of a springboard to start making a bigger RPG, which is something that I had tried and failed at years earlier.  I used game jams as self-leverage to make those games, but you don't have to.

There probably aren't very many comprehensive tutorials for more involved and specific games, but you should be able to find plenty for their individual components.  I'm sure there are plenty of guides for inventory systems, click-to-move controls, basic idle mechanics, various camera perspectives, and so on, which can be glued together into the the kinds of games you talked about.  A lot of your programming experience does help here, you just need some practice with the Unity layer.  For example, an inventory is just a data structure with a user interface; Unity's role would be in displaying the UI and detecting the user's input, so that you can translate that input into commands for the back-end data controller.

There are also some patterns and algorithms that are very commonly used in games, but not so much in, say, enterprise web development.  A* is one example that comes to mind, which is a basic pathfinding algorithm.  You're bound to come across those when you need them, so you'll pick them up over time.