Does the very fact that there is human variability void any validation that is carried out by anyone other than the operator used during the validation runs?
What you would really be validating in this case, is the training program used. So: You must show during validation that you have (previously to validation) trained the operators sufficiently to perform the assembly process within specifications (e.g. according to workmanship standards).
Alternatively you may validate the process controls or inspections used during assembly, if testing can be done using non-destructive methods. This may require less effort than validating the assembly process.
in order to fully validate the process would the qualification runs need to be executed with every Operator that might be assmbling for that process in the future?
After the initial validation, you should be able to introduce new operators using the same training program.
I would recommend that a risk based approach be used in evaluating the products assembled during the validation runs, so that inspection is tightist for assembly errors with high impact upon user risk.