Thank you all for the answers.
My idea is to spool the wire exactly from one "lip" of the coil former to the other without any gaps. To solve the gap problem, lag works quite nicely. However, even if I measure the wire at several place and use that measurement any slight variation of the wire creates problems when it comes to calculating the amount of turns required (coil_length / wire_diameter). Over several thousands of turns even a variation of 0.001mm can amount to a gross gap. I was thinking that actively measuring the wire diameter for discrepancies would allow me to set the number of turns as a "moving target" during coiling so that the wire ends up exactly at the lip.
Perhaps I should be looking for a strategy/some kind of sensor that would be able to sense the presence of the wire when it reaches the coil lip instead of trying to measure the error? I want to minimize contact with the wire as it is coated with a fine layer of plastic (that I want to avoid stripping away).