I'm not at all surprised to hear that there are organizations in India that take a more restrictive definition of engineer. The problem, I think, is that we're comparing a lenient count in India and China with a very restrictive count in the US. When you count more similarly, the gap vanishes - and actually looks more favorable for the US on a per-capita basis.

If you're interested, the Duke study that refutes the 350K number is available at: (warning, pdf!) http://memp.pratt.duke.edu/downloads/duke_outsourcing_2005.p...