Cortex: A Retrieval-Augmented Framework for Career-Aligned Learning Paths

Vatsala Ramachandran

doi:10.1007/s10994-026-07160-5

Cortex: A Retrieval-Augmented Framework for Career-Aligned Learning Paths

2026-09-22

0

PRE

AI

DOI：10.1007/s10994-026-07160-5

Original

Original request for help

Save

Abstract

En 中文

In an era of rapidly evolving professional development, many learners and individuals transitioning careers find it challenging to align their existing skills with the demands of their desired professions. Conventional career guidance systems often rely on rigid keyword matching, which overlooks the semantic nuances of users’ intentions. For instance, a user interested in “interpreting big data” may not explicitly mention “Data Scientist,” causing traditional systems to miss this connection. Furthermore, once a career path is identified, determining an optimal sequence for skill acquisition–while considering prerequisites and dependencies–remains a significant pedagogical challenge that is rarely automated effectively at scale. To address these challenges, this paper proposes Cortex, an AI-based career-aligned learning system that combines retrieval-augmented generation (RAG), large language models (LLMs), knowledge graphs, and sentence embeddings to semantically match user interests with occupations and generate prerequisite-consistent upskilling paths with inter-skill dependencies and scaffolding. The primary scientific contribution is a novel end-to-end architecture connecting semantic intent understanding, prerequisite-aware skill sequencing via ESCO dependency graphs, and taxonomy-constrained LLM generation. The system leverages the ESCO dataset to map career goals to relevant skills and produce structured learning paths. Experimental results demonstrate that integrating LLMs with a constrained RAG framework and knowledge graph-based dependency modeling can effectively generate prerequisite-aware learning paths for diverse career goals.

Keywords:

Learning path generation

Large language models

Retrieval-augmented generation

Knowledge graphs

Career recommendation systems

Cortex: A Retrieval-Augmented Framework for Career-Aligned Learning Paths

Abstract

Journal

Organization

Cited Papers