This is something I've never actually managed to answer myself, and I'm not sure how much of what I've done is right or wrong.

Say you've finished planning, and now you're ready to begin developing. In a medium/large project, what part of the game do you start on first? I usually start on the character & world design, but tend to run into quite a few issues/bugs, and get a bit lost in what I'm doing. I was thinking on our new project, to start on the User Interface first and then move onto world design, then character creation?

Thanks!

Normally I would informally aim for what is sometimes called a minimum viable product - i.e. what is the smallest program that does everything that you need to demonstrate the concept? Usually that means zero character and world design, and just about getting the technology in place to be able to show your game working with placeholder assets. That in turn means planning out the very high level feature set - e.g. "2D top down tile map with mouse interface" or "3D world with 1st person controls, mouse/WASD controls" and getting that up and running first, before digging deeper into game specifics.

It's funny that you say that you "finished planning" and ask "what to do next". When you have a GOOD project plan then you KNOW what should be done in what order. This is called "critical path".

I will agree with Kylotan with an addition. If you are introducing original game mechanics, you may want to try them first in the importance order.

For example, if you are making Portal, first you need to solve how to transport the player while keeping its momentum from one point to another on the map. Then you need to solve how to render "inside of the portal" because in the game, the player can see through the portal. Then solve how to make turrets see the player through the portal.

Because;

- the whole concept is about transporting the player through portals. You have to do it.

- It would be nice to see through the portals but you can still make them opaque, it would still work

- if the portals are opaque, the player can not see the turrets through the portals. So, enabling turrets see through portals would be a bad idea