Dynamically loading CUDA driver functions

Accelerated Computing CUDA CUDA Setup and Installation
1

For OpenGL there is GLEW, to load function pointers at runtime.

For Vulkan there is VOLK, to load function pointers at runtime.

I want to load CUDA driver functions the same way. I want to use nvrtc* and cu* functions. Is there such a loader library available? Reading the docs, to turn a string to usable functions I will need the following:

https://docs.nvidia.com/cuda/nvrtc/index.html#basic-usage
https://docs.nvidia.com/cuda/nvrtc/index.html#compilation
nvrtcCreateProgram    string   -> program
nvrtcCompileProgram   program  -> program
nvrtcGetPTX           program  -> ptx
nvrtcGetLoweredName   program  -> new function name

https://docs.nvidia.com/cuda/cuda-driver-api/group__CUDA__MODULE.html
cuModuleLoadData      ptx      -> module
cuModuleGetFunction   module   -> function

https://docs.nvidia.com/cuda/cuda-driver-api/group__CUDA__EXEC.html
cuLaunchKernel        function -> execute