I am doing a course in communications and while discussing multicaruer modulation to break down a signal into smaller bandwidths (for BW to be less than coherence bandwidth), there was mention of transforming the signal to frequeny domain before transmission. Logically, that doesn't seem to be useful to me as the hardware doesn't help us. The transmitted signals would get convoluted/multiplied as per the impulse response of the channel either ways and I can't seem to understand why a frequency domain transformation would help us. Is it something to do with reading the data for ourselves? Or does it improve SNR and SINR(Signal to noise ratio)?
At a very high level, multi-carrier takes a single-carrier communication system that would normally need a high symbol rate, and breaks it into many constituent signals, each with their own carrier and a much smaller symbol rate.
In the original single-carrier system, the symbol period may be much less than the length of the channel impulse response, so symbols get smeared and interfere with each other. But if you break that system into $M$ sub-carriers, then you can now get away with a symbol rate on each sub-carrier that is $M$ times smaller, or a symbol period that is $M$ times larger. If the smaller symbol period is larger than the channel impulse response length, inter-symbol interference is no longer a problem.
The DFT just offers a way to perform multi-carrier modulation efficiently, and the IDFT likewise offers a way to perform multi-carrier demodulation efficiently.
There is a lot more to it, but that's the basic idea.