Different systems and different people have different ways of solving the same issue... probably. What I am hoping to do is to try to do as much as possible using just 2D vector art, because it is lightweight, easy to animate, fast to render, fast to compute, etc. Basically, vector art for simplicity and speed, but in my tests, I am certain I will have to do a mix of vector and raster effects, as well as a layering and filter system. This is so the vector stuff can draw the clean lines and basic fill colors (as well as handle any occlusion mapping I need to worry about when different objects cover each other, ie, two characters hugging, on characters arm should cover the other's), while rasterization techniques (painting per pixel) can do some shading and coloring effects (redness for cheeks, shading for the body, etc) which I may see how I can port that to being done in shaders to help with computational efficiency.
I haven't played with unreal, so I don't know what language it uses. I like C# and its use of delegates to optimize on the fly behavior connection, but with the custom language I want to make, my goal is to make not a compiler, but a translator that converts simpler high-level code into higher speed lower level C code so that the code will have universal platform support, and should be run regardless of what language or computer/device the game engine actually uses.
To be honest, I am actually prototyping most of my tests in Blender, where I treat the vectors as 3D objects with shape keys, so the same code could be used with a 3D model with shape keys to generate the exact same body design behavior. In fact, it would probably be simpler to use a 3D model for this project as I wouldn't have to worry about shaders or 2D drawing techniques.
The reason I am doing 2D, and more specifically, the total drama series character style is, I have been trying to be a decent artist. I am more of a technical artist, meaning I like having the skill to make good art, but I don't actually have fun making art itself. I am more like you, electronics and coding are the things I really love to do. But I have dedicated a lot of time learning to do art, even spending two years to focus on human anatomy. I am ok with realistic bodies, but I still can not do stylized, cartoony, exaggerated bodies. The key to good 2d characters (my preferred art style is 2D) is being able to make the art easy to visually read and to exaggerate. this is because, without the third dimension, you lose that extra bit of detail, making it hard to read what it is you are seeing. a good example is to take a picture of a hot girl who is standing straight, then draw an outline of her only, the drawing will not be able to capture her charm without some shading or some exaggerated curves because we can not see the shapes in the third dimension. Exaggeration is the key to bringing life to 2D art, and the key to exaggeration is simplifying as much as possible. One technique is to use gesture lines. gesture lines can help describe a body type, its pose, as well as it can add to a basic action line (another mental tool used for drawing), if you were to draw a character, and paint it solid black, the gesture lines are the key to making a unique silhouette, and a silhouette is the first thing the brain understands when trying to identify what it sees. Looking around, the total drama characters are made up almost only of gesture lines, and by studying them I have a better understanding of how to represent a body type using gesture lines, how gesture lines can help create different types of body types, as well as some other tricks (like a curvy line should never have more than two nodes to control the curve of it, else there is too much detail for someone to pick up on). So I started this little project as a way of understanding how to draw cartoon characters, to see if I could use math to describe the shape of different body styles. if I can use math understand how to communicate body types just on gesture lines alone, then that means these two tools can be used together in code to generate a range of characters (poses and animating them as well) regardless of what type of art style the system wants to emulate. taking a big complex idea and breaking it down into its fundamental is the key to programming.
Hense, I am using 2D because I have made some simple mathematical observations, but you could just as well get a 3d body with some body-morphs aka shape-keys to design the body type you want, say in daz or whatever. I am just using 2d because I want to focus on a particular style/technique, as well as I want to make a robust animation system that can be adapted to any body type, as well as place some basic animations in typical of games. ie characters running for an rpg game. The idea is to be able to rapidly generate art for any of my future projects, be it a game, animation, or animated story.
