Another option would be wire spark erosion. It is a fair bit more expensive, but can do much deeper cuts and has no significant kerf.
I was going to suggest profiling with WSE twice, once at right-angles to the crank axis to get the waist, then the actual profile. However that would still require post-machining to make the round things round, so is probably not worth the bother.
This is clearly on a much smaller scale, but there might be some useful ideas for conrod machining.
http://www.f1-2000.co.uk/index.php?f=conrods