While I think the concentrated acid solution is brilliant, I'd like to suggest a refinement.
Rather than use just any old acid, you should specifically use a chemical etchant.
Typically, a chemical etchant's nucleophile strength does not come from Hydrogen, so it should have the benefit of much lower -if not entirely eliminated- hydrogen enbrittlement. (At least theoretically, anyway.)
A good etchant for this might be Ferric Chloride, [FeCl3]. Copper(II) Chloride [CuCl2] is another possibility.
Getting an even and precise etch will also be up to idiosyncrasy's of the exact setup you try and run. Chemical concentration, bath agitation, cleanliness of the work, temperature, ...
Another, related, option is to use electrochemistry instead of direct chemistry.
Specifically, use the reverse of electrowinning, electroetching.
One of the pros of this method is that, with controlled current + time, you could fairly easily calculate and control the exact mass of the material removed by the relationship between the unit of electric charge (coulombs), and the unit of molar mass (moles.) This extrapolates well to controlling the exact change in spring's mechanical property.
With direct chemistry, concentration + temperature + time = mass of material removed. Even if you precisely dial in the last two variables, the first one is going to be a ever changing target, as the work will dilute/poison the bath.
Probably the biggest con of this (aside from requiring a semi-sophisticated electronic power supply) is that surface irregularities will become much more of a problem than with chemical methods. In particular, any insulator will be practically inert to etching with the electrochemical method.
Conversely, an insulator might still be attacked by a pure chemical method. e.g. Any oil on the surface might become esterfied to something soluble, be freed from the surface, thus allowing the etchant to get at the actual work piece.