XR-CareerAssist: An Immersive AI-Powered Platform for Personalised Career Guidance Using Extended Reality and Multimodal Artificial Intelligence

Background Career guidance systems have traditionally relied on static, text-based interfaces that fail to engage users or provide personalised, data-driven insights. The rapid technological changes demand more sophisticated approaches that can help individuals navigate increasingly complex career landscapes. Computer-Assisted Career Guidance Systems (CACGS) have evolved considerably since their inception in the 1960s, yet most contemporary implementations remain constrained by limited interactivity and insufficient consideration of the narrative dimensions of career development. Methods This paper presents XR-CareerAssist, an innovative platform that integrates Extended Reality (XR) with multiple Artificial Intelligence (AI) components to deliver immersive, multilingual career guidance. The system leverages Automatic Speech Recognition (ASR) for voice-based interaction, Neural Machine Translation (NMT) for multilingual accessibility supporting English, Greek, French, and Italian, a Conversational Agent (Training Assistant) for personalised dialogue using the Langchain framework, Vision-Language (VL) models fine-tuned on career visualisations using BLIP with k-means clustering on representative Sankey diagram images, and Text-to-Speech (TTS) via AWS Polly for natural audio responses delivered through an interactive 3D avatar. Career trajectories are visualised through dynamic Sankey diagrams generated from a database of over 100,000 anonymised professional profiles. The platform was developed using Unity 2022.3 LTS with Meta SDK 2.0 for Meta Quest 3, with backend services deployed on AWS Elastic Beanstalk. A pilot study was conducted at the University of Exeter with 23 participants using structured questionnaires and semi-structured interviews. Results The pilot study demonstrated strong validation results including 95.6% voice recognition accuracy, 78.3% overall user satisfaction, and 91.3% positive ratings for system responsiveness. The platform architecture employs AWS cloud infrastructure with FastAPI microservices, achieving response times under 200 milliseconds for career map generation. Load testing validated scalability to 10,000 concurrent users with 900-1,000 requests per second and zero failures. Critical user feedback led to 100% implementation of improvements addressing motion comfort, audio clarity, physical boundaries, and text readability. Conclusions XR-CareerAssist represents a significant advancement in career guidance technology, demonstrating how the convergence of XR and AI can create more engaging, accessible, and effective career development tools. The successful integration of five AI components within an immersive XR environment creates a multimodal interaction experience that distinguishes XR-CareerAssist from existing career guidance systems. A preliminary version of this work was presented at the XR Salento 2025 International Conference on Extended Reality. 1