On second thought, inverting the assembly would be a bad idea if the goal as to gain control over the angle of the bit. You’d want the point of tilt to be as close to the spindle end of the router as possible, to minimize the amount of X, Y and Z shift at the cutting surface as a part of the angle change. You could still do it inverted, and it would add less material to the sled for a given amount of angular pitch, but there would need to be (more) compensation in the software, and you’d run more risk of hitting the router body on the work surface at the extreme ends.
Edit # I do this a lot: My little doodle up there isn’t really right, the arm would need to mount somewhere on the middle of the adjustment screw when at 0 angle deflection, to give room for the armature mount to shift down during angle changes.
Still more edits: The way delta printers are set up, two struts connecting a bracket on the height adjustment to the print head with fixed brackets on both sides, allows the print head to move while keeping the print head level to the print surface. To allow for angle shift you would either need an extra joint on one side of the connecting arm, or maybe some sort of Y shaped connecting arm with a full ball and socket on the single connection point side, single connection points on both ends would lead to twisting issues I think.