Simple CUDA Code Cython Integration

I am looking for an experienced developer who can use Cython to wrap my existing CUDA code. This is a relatively simple code that is currently used for data analysis. Task: This is done in Linux (ubuntu). I had a old cython code that used to be able to connect with the CUDA code, i think the newer version of CMake changes had made it impossible. I have attached a small reduction of the code, it is pretty much just calling a CUDA kernel from the Python using Cython. The CMake builds, but when the Cython is imported the symbol cannot be found. (to build you need to define an environment variable CUDAHOME points to your nvcc) Steps: 0 (optional) you can Make the cuda code inside to check if it works (1) run [login to view URL] (with CUDAHOME env var) (2) run [login to view URL] (fail at this point as the .so file generated does not contain the symbol (undefined symbol)) Ideal Skills: - Proficient in both CUDA and Cython - Experience in wrapping CUDA code using Cython - In-depth knowledge on data analysis algorithms - Ability to deliver high quality, optimised code Your main task will be to ensure seamless integration between the Cython and CUDA languages, ultimately elevating the efficiency and performance of my existing data analysis algorithm. Only those with relevant experience and skills need apply.