Numpy/CUDA/Python Project

I am looking for an expert-level, reliable numpy developer who can start an existing python project (that already uses numpy) but the components needs to be made much more capable and better performing. Pandas and/or Numba ok. A detailed algorithm spec exists. I prefer to have it running on a GPU/CUDA. The algorithms can fundamentally be parallelized into completely independent tasks. This project requires a close collaboration with me and someone with good English communication skills. Experience in architectural work is also being required; good math skills and data science experience is definitely a plus. I am looking for someone who can start immediately and can dedicate his fully time/attention to this project until it is done. There will be several follow-up projects if I can see that we can work together. For this projects, python 3.6 code has already be developed; I will share documents with qualified developers. Also, if you don't have the time/ability to dedicate your full time to this project until it is done, then please don't apply. After the project is done I will need from you the written confirmation that with my final payment I have received the entire intellectual property within this work for hire project.