Hello,
We have already talked about this, I am just updating my bid now because I read the paper on Keeloq you sent. In case you are still interested we can talk further for details
I am very i interested in this kind of low-level GPU programming as I have familiarity of the parallelism it involves, previously I have optimized Turbo Decoding in GPUs which was a project that required knowledge in such techniques (bit level and thread level). I have also used distributed computing but not in the AWS you want it to be deployed. Anyway it does not seem to be difficult task.
In case you reply I will have some more questions, mainly on the financial part of this!