Since the interaction between wires in a transformer is by photons and not by this fictitious force of magnetism the electrons in one wire connect to the other by the photons emitted as the velocity of individual electrons change.

This means we do not need the heavy Iron laminations, the connection is electrostatic in nature (by photons) so the iron should be between the wires that are interacting, this applies to self-inductance as well so the best results will be by putting dust iron in resin and embedding the wires of both windings by potting in this resin containing dust iron.  Or use a low loss ferrite.

Because the permeability is related to the permittivity of an dielectric by the velocity of light then a dielectric may be used but the value is much smaller. u=e/(c^2) here u is the permeability and e the permittivity.

Possible design of an inductor

Transformer or mutual inductor design.  A simple inductor is similar but with only one winding.

Another improvement is a copper screen round the outer winding and another one round the inner winding to act as a screen against photons escaping thus reducing losses still further.