Hi Sylvian - I would get all the ‘obvious’ views of the object if possible - side, top, etc. Any drawings that look reliable would help as well. I’d place the images in Rhino using BackgroundBitmap, or PictureFrame.
Then, and this is the tricky part if you are just getting going, I would spend some time thinking about how to organize the surfaces - this is not an especially easy project to do well, so you might want to familiarize yourself with Rhino’s various surfacing tools with some simpler projects. As a general strategy though, it’s usually useful split up the shapes mentally into ‘primary’ and ‘transitional’ surfaces and model them separately . So for example, you might want to make the red object at the right as a hard edged ‘wedge’ shape with simple, clean surfaces and sharp corners, and then make the blends or fillets that soften it afterwards. In short, keep your model organized as much as possible and try not to force a single surface to accomplish too much… the result will tend to be crisper and the design more readable.