Utilizing Deep Learning for Semi-Automatic Conversation Analysis During Recruitment and Employee Education in the Seed Phase of High-Tech Startups

In high-tech startups in the seed stage, grappling with limited staffing resources, the expeditious discernment of an individual's potential as a pivotal element of the institutional structure is of paramount importance. The present research enabled a substantial diminution in evaluation duration in contrast to traditional methods by harnessing the power of Deep Learning for transcription, automatic speaker differentiation, and extraction of speech attributes during the Deep Learning procedure. Furthermore, utilizing the attributes excavated via Deep Learning for textual analysis through Principal Component Analysis (PCA), it became feasible to glean quantitative propositions from a broader range of perspectives hitherto unobtainable. Despite the concentration of this study on attributes pertinent to dialogic velocity, it elucidates the feasibility of accruing more multifaceted suggestions through the extraction of additional attributes, such as vocal pitch, from the internal output of Deep Learning and associating them with analysis findings of PCA.