I feel this terminology is not well defined and it is difficult to tell what it means out of context. What I believe your professor had in mind, relates to how many options are already "built in" in the system and are available for the user-programmer. I will explain.
Let's say you are engineering your own embedded card. You pick a CPU. You pick controllers. You decide the interfaces to the outer world. Etc. Then, once the card is manufactured, you will have to write the drivers by yourself - the pieces of software that allow the CPU to operate the system and mainly operate the external interfaces of your card. This would be a "bare metal" system, as it provides you (the programmer) no pre-defined drivers or operations.
On the other hand, you could base your system on a pre-designed micro-controller. Such a chip already includes a cpu along with many peripherals that interface the chip to the outer world (cf. a system on a chip). Operating this microcontroller chip is quite easy as the "hard work" was already done by the designers of the chip. Complex operations, e.g., converting digital information to analog information to be output by the analog output, are as easy as writing a byte to some pre-defined address. The microcontroller will know to take the information from that address and do all the rest. That is, the designers of that chip provide you with a "fully featured", easy to operate chip.