After investigating with a bit more Dykem, I don't think the two points in the corner of the lever are causing the BCG bind. Perhaps relieving those areas on the lever would make the lever bind even worse, as would relieving the aft end of the upper slot (where the perimeter of the slot has a recessed edge) as some SS manufacturers have recommended and some users have done.
As Tim suggested, the bind is between the inside BCG surface of the upper and the BCG so relieving the areas described and illustrated above in my photo might allow the lever travel further rearward, where it might bind worse.
So removing some material from the underside of the lever's tail, as Tim also suggested, seems to be the best fix. I'd already removed some material from the topside of the tail but I now suspect there's not a lot of extra height available on the top, so the underside is probably the best place to "thin the tail."
I've now removed a ton of material from the underside of the tail...well, at least 1/3 of the tail thickness documented by the Lever V4.4.STL model file, and the metal lever had quite a bit more thickness than the .STL / nylon printed lever to begin with, so there's no wonder the metal lever jammed! The nylon lever is slippery Taulman 680 (also quite stiff with a high transition temperature) so it doesn't cause a bind at full model thickness like the metal lever.
After countless filing and grinding iterations, there's now no binding with the metal lever and the bolt catch works with three quarters in the buffer tube. I'll thin the lever tail more and, hopefully, get it down to zero quarters before I break the SS with too much lever thinning!