Just melting things makes a planet with an iron core and a silicate ("slag") mantle and crust, because Iron is denser than silicates. It would be just like our earth, but a bit more baked. (like the Moon, in fact, but with a bigger core).

You have to vaporize something to get rid of it from a whole planet. For small particles, it is the iron that vaporizes first. refractory silicates are the last thing to go, and the first thing to form in a planetary nebula (Spinels and other ferro-magnesian silicates form at high temperatures from vapour. So does Diamond, btw.

But we're running things the other way. We're taking an EXISTING planet and heating the surface.

So, you boil off the ocean and strip off the atmosphere first, then melt the surface to lava, then start boiling away the surface of a magma ocean. (which can still be solid at depth because pressure encourages minerals to solidify).

To strip-away a LOT of silicates, you would need a VERY hot silicate vapour atmosphere for the heavy silicate elements to reach escape velocity. So i don't think you CAN, in practice, strip away significant amounts of silicate crust and mantle. (but you can boil out the volatiles, and maybe some of of the lighter volatile metals like lithium or even sodium)

It also depends how fast you heat it, and how long you heat it for. If there were no internal sources of heat, it takes about 1 million years for a thermal wave to penetrate 4 km through rock. So you'd need ~1 billion years to work down through Earth's 3,500 km Mantle and finally start on the nickel-iron core.

I don't know, off-hand, how long a red giant phase lasts, but 1 billion years is possible. At the end of this time you'd have a fully-molten and volatile depleted ball of magma with a liquid iron core.

But there is Internal heat as well. Over a billion-year time frame, this will stir the mantle and core by convection and help melt them from inside-out, accellerating the initial melt-down, deepening the magma ocean, and making everything a little quicker.