Design of personalized career planning assistant based on large language models

Scientific career planning helps individuals clarify development goals, enhance competitiveness, and align personal growth with industry needs. However, due to the scarcity of certified career counselors, many young people face challenges such as unclear positioning and vague development paths. Existing career recommendation systems often rely on static rules and lack personalization and dynamic feedback, limiting their accuracy and diversity. To address these issues, this paper proposes a personalized career planning system based on large language models (LLMs) and semantic retrieval. The system constructs dynamic user profiles by extracting interests, skills, values, and personality traits through multi-turn dialogue, incorporating the Big Five personality model for psychological profiling. It applies a Hexagonal Competency Model to structure job requirements and uses SBERT embeddings with FAISS for precise vector-based matching. Personalized recommendation explanations are generated through LLM prompting to improve transparency and user trust. A user survey showed that 83.1% of participants rated the recommendations as fairly or highly matched, demonstrating the system’s effectiveness as a scalable, interpretable, and practical solution for career development, particularly for university graduates.