The gameplay video gifs might take a bit to load, sorry :)
Footage credit to GAMEPLAYBOX on Youtube.
I created and worked on 200+ levels, and took levels from the idea phase all the way to production ready phase.
I participated in ideation meetings, helping to create and flesh out puzzle mechanics.
I finalized the visuals for tons of levels, shaping the platforms and decorating the scenes.
I proposed an experience design feature, which was implemented by the developers.
I made loads of bug reports while testing levels, providing videos and descriptions of how they can be reproduced.
The established design rules were to make short levels, preferrably taking only one retry, with the least possible different types of mechanics present per level (max 3).
To make interesting levels I started by thinking of fun movements/interactions that are possible with the mechanics. If I couldn't think of any, I'd arrange mechanic objects in random or fun looking ways in the Unity editor, and experiment around to see if I could create fun movements/interactions from those arrangements.
In all levels, players were supposed to be able to finish with the first 2 stars easily, the 3rd one being the big challenge. The core puzzle interaction designed above in step 1 is mostly the path to get that 3rd star, making the puzzle solution always a well thought out movement/interaction.
I then designed a path that only got you the first 2 stars, but in a way where that seems like the most obvious and easy choice. For example by placing near the player's start, making you think that's the first one to get, but when doing so it will deny the 3 star path. I also placed and shaped all level objects in ways that they guide you towards the 2 star path.
After the first 2 steps I had a of puzzle with hopefully a statisfying solution. With this method I created batches of 10-20 levels that I playtested with the designers and sometimes on playtest days with other team members.
The designers would then give me feedback about how to fix or improve the level. Eventually they would either approve each level or explain to me why the level's core puzzle isn't good enough or doesn't work, after which I would then drop that level and write down the feedback for future levels.
Finally, we took all approved and selected levels, and polished the look and gameplay experience to make them production ready!