Navigating the pressures of job interviews can present a significant challenge to an individual as their preparation can critically influence the outcome. While traditional face-to-face (F2F) methods for interview training are still the norm, the evolution of ‘artificial intelligence’ (AI) presents new opportunities, and a variety of virtual AI-based interview platforms exist to satisfy the need. However, while such platforms offer a wide range of features and broad coverage of interview-style questions, they often lack the physical presence and the capacity for real-time emotional feedback or adaptive response tailoring that is characteristic of real-world F2F scenarios. As a consequence, there is growing interest in the use of physical robots for job interview simulation. This paper presents the development of an advanced AI-based ‘Interactive Interview Training System’ (IITS) utilising the FurHat robot which goes beyond the feature sets of existing virtual and physical agents to create a more holistic and effective job interview simulation. In particular, FurHat has been configured to allow extensive customisation by the user, such as the ability to define job scenarios from different industry backgrounds and invoke a range of interaction styles. Also, since FurHat’s Software Development Kit (SDK) incorporates a virtual animation of the robot, it has been possible to conduct a user-based evaluation which confirmed the merits of physical embodiment in HRI-based interview training.