I am a quantititative analyst, specialized on simulation intensive computational problems. For that I want to have an automatic (script-based) provisioning of an MPI Cluster, on EC2 or GCE. On these MPI nodes I want to run my libraries which use Python/C++/Cuda. The libraries build on my custom docker image, hence I guess containerized platforms like Kubernetes, Juju or else can be interesting to use (but I am open to hear about your proposals).
Therefore, I would expect to work with a DevOps/HPC professional that can provide me with a working scripted solution:
- to deploy automatically an MPI cluster on the cloud, with my library dependencies or my docker image.
- to run and debug from my laptop or any other PC against the cluster
I am happy to provide more info, if it is required.