What other types of layers ? Those I mentionned at the end like Convolutional or Flatten ?

In general, the types of layers that you use (which at the end makes up the architecture of your neural network) are really data dependant. Convolutional layers for instance, are simply more suited for images than Fully Connected layers are.

Other layers like Dropout are purely practical and can be used in any network. The Dropout layer randomly sets to 0 some weights or biases at each pass so that the network learns to output the right result without going through the same connections every time. Dropout helps to reduce overfitting.

Many others exist, you should find all the information you need if you do some research.