Speech quality assessment in the filed of machine learning is associated with setting up an automated system that is capable of characterising the quality of human speech in terms of factors such as the expressiveness and contagiousness of the speaker. Such a tool can have many practical applications such as the self-improvement of the skills of a public speaker and the detection of learning difficulties (eg dyslexia). This internship focuses on gathering and annotating of english data related to speech quality in order to upgrade the developed pipeline of automatic speech quality assessment (https://github.com/tyiannak/readys). Specifically the labels of interest are the following: expressiveness, ease of following the speech and enjoyment. Another task of interest is designing and implementing improvements to the system (eg upgrading internal/final models, adding meaningful hand crafted features). Possible investigation of the use of the pipeline in solving other problems related to speech (eg. engagement and empathy) are also in demand.