CareerPath AI: Predicting Student Career Domains Using Academic Performance and Skills

Abstract

Picking a career is a big deal, for students. It is one of the important choices they will make. When a student chooses a career path it can affect the rest of their life. Choosing a career path is something that students should think about carefully. Yet career counselling at colleges is not good enough in three main ways. Counsellors do not have time to give every student personal help. The advice given is often not specific to the students schoolwork. Also many schools do not have a system, for giving career help. This paper presents CareerPath AI, a system that looks at both a student's academic performance and their skills and interests together to suggest a suitable career domain. The system starts by cleaning and arranging the raw student data. Then it takes out features. After that it trains a prediction model. A number of machine learning algorithms were looked at. These algorithms include Logistic Regression, Decision Tree, Random Forest, Support Vector Machine (SVM) and XGBoost. The goal was to find out which one predicts career domains best. Based on this, the system outputs a predicted career domain for each student, such as Data Science or Software Development. The paper also looks at which features matter most, how to measure the system's performance (accuracy, precision, recall, and F1-score), and what it would take to actually use a system like this in a real institution. Finally, the paper is upfront about the limitations of this approach and suggests future improvements, such as making the model explain its predictions more clearly and adding psychometric data.