CUDA Developer/Coder Required To Optimise Code In A CUDA App

I have a cuda based app that uses the cuda cores on nvidia graphics cards. The code was written about 6 or 7 years ago and is not optimised for todays generation of graphics cards and PC hardware.

I think the code can be improved to run more efficiently and faster, this is the need for the project.

The CUDA developer/coder needs to have an understanding of graphics card architecture and be able to look at the code of the existing app and improve that code so that it runs a lot faster.