Radio and optical telescopes observe the radio and optical portions of the electromagnetic spectrum correspondingly. But radio telescopes are fundamentally different from their optical counterparts. Radio telescopes don't have CCDs. They don't need finely polished mirrors and they are much, much bigger in size. So. Why the fundamental difference?
Let's take a detour and remind ourselves of the wave-particle duality of light. The popular young's double slit experiment is explained using the wave theory of light where as photo-ionization is explained using the particle theory of light i.e using photons.
A wave can be described using y = A cos( omega*t + phi ) where omega is two pi times the frequency, phi is the phase and A is the amplitude of the wave. A photon on the other hand can be described using h*nu where nu is the frequency of light. This fundamental difference is what, as you will see in a moment, lead to the differences between optical and radio telescopes.
Optical detectors such as CCDs are designed to capture the photons I.e the photon is absorbed by the CCD and its energy is therefore registered. The intensity of radiation is understood from the total number of photons captured by the CCD in one unit of time.
On the other hand, radio telescopes record the wave nature of light I.e they record the A - amplitude of the wave, omega - two pi times the frequency of the wave and phi - the phase of the wave. The last bit, the phase of the wave, is what makes all the difference in the world. Optical detectors lose information regarding the phase of a photon but radio detectors don't.
Recording the phase of waves can help study the polarization of light and the phase information helps in interferometry, something very hard to do with optical light. Special instruments are required to study polarization of light in the optical domain but it is trivial with radio detectors. And interferometry is simply adding waves from different telescopes in phase, which given the reasons mentioned above, is not easy with optical telescopes but is trivial with radio telescopes. Don't get me wrong, there are optical interferometers, such as the Large Binocular Telescope, i'm just saying that it's harder to do in optical than it is in radio.
Interferometry is infact necessary in the radio domain. The diffraction limit puts a limit on the angular resolution of a telescope given the wavelength of observation and the size of the primary mirror. Since radio waves have wavelength many thousands of times that of optical light, individual radio telescopes have bad angular resolution. For example, radio waves have wavelength of the order of a few mm to cm where as optical light has wavelength of the order of micrometers. Again, interferometry comes to the rescue. Once two radio telescopes are linked together for interferometry, their individual primary mirror sizes don't matter anymore. It is the distance between them that matters, referred to as the baseline. You might've come across the phrase 'Very Long Baseline Interferometry' which originates from the use of radio telescopes separated by very large distances. Interferometry therefore helps a radio astronomer get angular resolutions similar to what an optical astronomer would while getting information on the phase of the incoming radiation in addition to the slew of other properties mentioned above.
Moving one step ahead, aperture synthesis is trivial in radio astronomy. As mentioned earlier, given two radio telescopes, their primary mirror size doesn't matter as much as their separation does. The primary mirror still matters because it decides how much signal we finally receive. Therefore, the space between the longest baseline can be filled with more telescopes, separated by various baselines, all of which work together to give a well resolved and strong output. The GMRT, located in Maharashtra, India is an example of this.
One last item in the list. If you have noticed any of the pictures of radio telescopes, the primary mirror isn't as shiny as a primary mirror of an optical telescope is. Surface roughness is a measure of how rough or smooth the surface of a primary mirror is and it in turn effects the angular resolution of a telescope. It is suggested that the surface roughness be of the primary mirror be of the order of lambda/20. In case of optical telescopes, this amounts to a few tens of nanometers assuming optical light of 500 nm whereas in the case of radio telescopes operating at 2 cm, the surface roughness can be of the order of a few mm. Light weight aluminium panels with surface roughness of the order of a few mm are used. This is one of the main reason why the primary mirrors of steerable(!) radio telescopes can be as large as 40m, while still being light in weight whereas the largest in optical telescopes is 10.4m beyond which weight of the primary mirror causes problems for itself.