In my computer vision startup, we just got to the point where running datasets on a single development machine takes hours. This hampers progress, since one needs to know performance metrics before making changes to the product.
We're looking into how to make this process faster by distributing the workload across many machines.
My question is, given a job (e.g. run a program on a dataset and collect some results), what are some practical frameworks that are out there that one can configure and use with relative ease. Note that none of us is a cloud expert but we are very comfortable with linux.
thanks in advance.