Welcome to Sound Design.
Your co-host will need to use headphones/earpods, etc to stop your voice interfering with their voice.
You can use a TRRS splitter to combine the microphone and headphone leads. There are plenty around but as you're using a Rode mic they make the SC6 for TRRS inputs or the SC6L for lightning inputs.
This is one reason why a) you do a level check before you go & b) why you record to separate tracks.
If compression/limiting can't make up the gain enough, or over-compresses one voice without making enough gain in the other, then you're probably going to have to manually chop into segments & move one voice to another track, so you can then treat ...
Upsample them all to 24 or 32-bit 48kHz* WAV at or before import.
This will gain no 'new' quality but will enable more manipulation/plugins etc without losing anything further.
You then have the highest quality master you can achieve, which you can downsample afterwards to whatever broadcast structure you need.
*or 44.1 if you don't need video sync.