AIPCRS: An Interpretable and Resource-Efficient Career Recommender Using TF–IDF Ranking and Decoupled Machine-Learning Microservices

Selecting an appropriate career pathway requires students and early-career professionals to relate heterogeneous skills, education, experience, career objectives, and work preferences to increasingly diverse occupational information. This paper presents AIPCRS, an interpretable career recommendation system that formulates candidate–job matching as a multi-attribute profile-to-document ranking problem. Career goals, skills, preferred job information, educational and professional attributes, and location preferences are combined into a composite textual profile and represented using term frequency–inverse document frequency (TF–IDF), after which cosine similarity ranks candidate job documents. The recommendation engine is deployed as an independent Python/Flask microservice communicating through a REST interface with a React, Node.js/Express.js, and MongoDB application. Prototype verification comprised four recommendation-engine unit tests and five end-to-end functional scenarios, all of which produced the expected behaviour. Illustrative recommendation analysis further demonstrated differentiated ranking across occupational alternatives. The study establishes the technical feasibility of an interpretable and modular career-recommendation architecture while identifying semantic representation, labour-market freshness, fairness, external validation, and production deployment as priorities for subsequent investigation.