I'd say start looking for members once the core features are implemented. And yes, that means that at the beginning you'll have to work all alone. That's the nature of the matter.
You will get team members later, if the first alpha and beta builds are convincing. Believe me; you don't want user contribution of inexperienced modders anyway, so advertising this before at least an early alpha of the gameplay is out is pretty much pointless.
Working alone for the first part of the map developement will teach you the bite and dedication you need to finish the project alone. And in the end, you will always be alone. Team members will only get you so far. The core work will always be done by the project leader, which means that you basicly have to know about every aspect:
terraining, modelling, scripting. If you lack in one section, go study that. You can ditch the modelling at first and learn about that later, but terraining and scripting are important from the get-go; otherwise you will have to completely makeover your game several times (been there, done that).
If you haven't already, start with a piece of paper. Write down exactly what features you want to have in your game. It helps structuring the map and keeping the content clean of inconsistency. Don't try to have too many features in your map at first. Content trumps features in terms of RPGs.
Avoid spaghetti code and hardcoding things when starting out with the game. Make it as modular as possible. Write registries for every aspect of content. Design your content to be as dynamic as possible.
If you plan with 10 classes at first, make sure you don't base all your systems on those 10 classes. Make sure that with just a few lines of code, you can register a new class to your system without the need to search through the whole map script, etc.. Again, been there, done that.
Btw, starting out with the inventory system is probably not the best approach. You should focus on the core game mechanics at the beginning, those are:
1) creep respawn system
2) hero selection (with placeholder heroes... start with a very low number of classes at first and then expand later)
3) hero progression (write the systems of customization; skill learning, ability learning, you name it)
4) experience and gold gain (you'll probably want a dedicated system for that)
5) a creep AI system (unless you only want to have object editor spells, you will need one)
6) a spellcasting system for better control (you might want to modify the way wc3 spells work, like visual casting time, dynamic mana cost, etc.)
7) writing a custom buff system to make spell coding easier (though I really recommend to use one of the existing libraries)
8) a robust creep camp engine that can not be exploited (the default WC3 AI does not handle this well enough)
9) a robust and easy to manage stat system, depending on what you want to have.
Inventory systems are just graphical stuff and have a much lower priority than this.
