A systematic review of skill-centric career guidance systems: Integrating knowledge graphs, skill decay tracking, and country-based salary sequence intelligence

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

With the rise of online learning sites, professional networking platforms, and hiring services, the demand for effective systems to support skill development and career advice is increasing. Skills are stored on different resources like e-learning platforms, code repositories and professional profiles. In recent years, personalization has been an active research field, and a personalized learning recommendation system, which recommends resources according to users' choices, has been developed. However, they have difficulty accommodating shifts in the labour market and are prone to creating negative employment mismatches and limited alternative career opportunities. The paper provides a comprehensive review of the research conducted across skill extraction, large language model (LLM)-based career guidance, knowledge graph construction, learning path recommendation, prerequisite modelling, and knowledge tracing, fairness and algorithmic bias in career AI, and country-based salary integration to improve adaptive career guidance. Almost 45 peer-reviewed articles from 2018 - 2026 were collated from key digital libraries. The review highlights progress in areas such as skill inference, learning path recommendations using graphs, knowledge tracing that considers forgetting, career matching, recommendations that explain themselves, and understanding salary trends at a country level, which is important for career planning systems. However, these solutions are not integrated and do not create a system for learning altogether. They don't consider the monetary aspects of paying for a career either. The review also proposes a series of skills and financial outcomes to help learners identify if they have the financial resources and location to pursue a career. It highlights the need for a platform to enable multi-platform skill extraction, tracking of skill decay, dynamic career paths, explainable AI skill recommendations, prerequisite modelling, and geographic salary return sequences.