I would submit a detail of the "corrective" action taken to the Engineer to keep him in the loop. I agree that what has been provided may surpass what was specified by the structural drawing, but the Engineer is responsible for reviewing and approving all joint details once the fabricator has completed the detailed shop and erection drawings (refer to D1.1-2008, clause 1.4.1, 4th sentence). I don't believe the Engineer will have any heart burn over the change in detail since it doesn't involve a reduction in allowable stress. The shear on a CJP groove weld is the same as for a fillet weld (refer to table 2.3).
At the very least, you need to provide the Engineer with data supporting the fact that the resulting weld is at least as good as that originally specified by the detailed drawings. What you described, a backing weld, followed by back gouging and welding should be sufficient to satisfy the Engineer that the weld is at least as good as the original fillet weld specified. However, it may be advantageous to provide the Engineer with UT data to confirm the back gouge operation and subsequent welding does meet the acceptance criteria for CJP.
Keeping the Engineer in the loop is required by D1.1 and it is good politics. Engineers know fabrication errors are going to happen. It eases their mind if the fabricator provides them with a "fix" and the necessary information, i.e., calculations, to make an informed decision whether the "fix" is adequate for the intended service and load conditions. After all, corrective work is the responsibility of the fabricator, but the Engineer must approve all repairs and corrective action. It is allways bad politics to wait until the third party reports the nonconformance.
Best regards - Al