First, calculate the free energies of reaction for all reasonable reactions you can expect in the system, including reactions of your nitride(s) with the molten salt. If the delta G of reaction for your desired reaction is significantly positive you're out of luck and it's not worth putting in the effort.

Generally the free energy of formation is more negative for oxides than for nitrides, as you probably know, so if you start with oxides you are not likely to be able to form nitrides. But there may be some exceptions so it's worth inspecting the available literature.

A more likely to succeed strategy might be to start with binary nitrides and try to get them to react with each other in a molten salt medium, making sure that the nitrides do not react with the molten salt itself. If you do this, make sure to remove oxides from all starting materials and oxygen from the atmosphere under which the reaction is performed, because otherwise your product is likely to contain oxides or oxynitrides.