I did it a little differently for the top layer on the splatch version of the bigger experimental sensor (shown in the photo above). I include that here, since it might be easier to understand how it relates (see attached). In total I now have sent three different experimental nodes (small, medium, and large) to that fab that have been revised to include PCB traces and pads for connecting up a micro splatch. On the big experimental node (shown in the attachment to this post), you can see them at the very bottom. The datasheet doesn't mention anything about the minimum width for the ground traces, but I beefed some of them up a bit. I didn't use a copper pour on the earlier experimental nodes, and even the small one seems to be working just fine without it.