Produce individual figures for each of the profiling runs

[max-models]

And push them to the profiling-data repo