Today, the discovery and development of new medicine is very costly due to the long research and approval time, typically ranging between 10 and 20 years. This long discovery to market time is a consequence of a very complicated search for a specific functional behaviour as the functional molecules constitute a very small subspace within an enormous chemical search space. For instance, a small protein with a length of 30 amino acids resides in a space of 10^39 candidates (and that is only considering natural amino acids). Furthermore, the average research and development cost for each successful drug is estimated to be around 2.6 billion USD – including cost of failures. In order to explore the high dimensional design space more efficiently and to reduce the number of design cycles, and therefore save both time and resources, it is pivotal to research deep generative models that would allow for generation of the qualified compounds automatically.