To make sure I'm happy with the overall geometry, I completed the design, including loops for elastic.  The fit is rather good on many adult faces (although I know I'll ultimately need different sizes).

The next step is CNC routing an MDF mold for the vacuum former.  Since the part is 3.5" (~90 mm) tall, I needed to order some unusually-long end mills to avoid crashing the machine; I picked up a pair of extended 7/16" square-end mills and a pair of 1/4" ball-end mills (chosen largely based on what was cheap on ebay).  I'm still teaching myself 'proper' CAM in Fusion 360, but the resulting part came out pretty good:

Next step was to our ancient vacuum former (whose manufacture was probably closer to the Spanish Flu than the current Coronavirus outbreak...), where I had fairly good results from 0.030" (0.75 mm) polystyrene. 

I'm still fighting some annoying issues with seams.  It would appear that my most recent pass is my most successful, where I had the idea of reducing the heating time, making the plastic more viscous and less likely to pinch easily.  Another approach would be to change the angle of the nose cone section, from it's current 10 degrees (if I recall correctly) to something more aggressively chamfered.

Last step was to the laser cutter, where my 4" (100mm) focal length lens was just barely long enough to cut the strap slots and outside profile of the mask, liberating in from the remainder of the sheet:

Not too bad! Next step is to try to meet up with some researchers here in Tucson at the University of Arizona, who have a respirator efficiency testing rig, to determine how well it performs versus existing commercial designs.  Fingers crossed!