AI-Enabled Personalized Interview Coach in Rural India

Abstract

Soft skills are one of the highly-valued skills integral in one’s personal as well as professional life. Recent advancements in technology and the penetration of the internet enable new ways to build systems to identify and help learners who need assistance in improving their soft skills. This paper presents the design, development and evaluation of a prototype intelligent system for training youth in rural areas. The system assesses a personal interview of a learner using their audio/video captured from a mobile phone or web camera. Verbal and non-verbal cues are evaluated using speech processing, computer vision and natural language processing and a detailed feedback is generated using Natural language generation. The real-world data used for training the machine learning models for the system were collected from rural areas in India and were annotated by reliable experts. We validated the performance of the system on an out-of-sample data-set which included 100 interviews collected by piloting the application.