Smart Career Guidance Platform Using AI for Career Mapping and Job Role Prediction

Abstract

Choosing an appropriate career path is among the most consequential decisions in a student’s academic life, yet it remains a challenging endeavour owing to the lack of structured guidance, rapidly evolving job market dynamics, and limited awareness of requisite skills across domains. Conventional career counseling approaches often fail to deliver personalized recommendations, leading to suboptimal career choices and academic disengagement. This paper presents a Smart Career Guidance Platform built upon Artificial Intelligence, designed to assist students in identifying suitable career paths by analyzing their skills, interests, and behavioral preferences. The system employs a two-stage machine learning pipeline: a Random Forest classifier predicts the most appropriate career domain from user survey responses, while a LightGBM classifier recommends the top three relevant job roles based on the user’s technical skill profile. Feature selection techniques including Correlation Analysis, Principal Component Analysis (PCA), and SHAP (SHapley Additive exPlanations) are employed to enhance prediction accuracy and model interpretability. The platform additionally generates a personalized career roadmap providing step-by-step guidance for skill development and career growth. Implemented as a Streamlit-based web application, the system achieves a prediction accuracy of approximately 90% for career field classification, demonstrating the effectiveness of AI-driven decision-support systems in the domain of career guidance.

Introduction

This paper presents a Smart Career Guidance Platform that uses Artificial Intelligence (AI) and Machine Learning (ML) to help students and early-career professionals make informed career decisions. Traditional career counseling methods often rely on manual guidance, aptitude tests, and static information, which fail to account for the diverse combination of skills, interests, behaviors, and rapidly changing job market demands. To overcome these limitations, the proposed system provides personalized, data-driven career recommendations through an interactive web-based platform.

The platform employs a two-stage machine learning pipeline. First, a Random Forest classifier predicts the most suitable career domain based on a user’s interests, preferences, skills, and behavioral traits. Second, a LightGBM model recommends the top three job roles within the predicted domain. In addition, the system generates a personalized career roadmap that outlines learning phases, key topics, and relevant certifications to help users achieve their career goals.

Objectives

The main objectives of the system are to:

• Predict suitable career fields accurately using AI.
• Analyze users’ skills, interests, and preferences.
• Recommend the top three job roles matching the user profile.
• Generate a structured career development roadmap.
• Deliver recommendations through an easy-to-use Streamlit web interface.

Literature Review

The review highlights the evolution of career guidance systems from rule-based approaches to AI-driven platforms. Earlier theories such as Holland’s Career Choice Theory and Social Cognitive Career Theory provided valuable conceptual frameworks but lacked computational adaptability. Traditional machine learning methods such as KNN and Naive Bayes have been applied to career prediction but face limitations in scalability and accuracy. Recent studies demonstrate that AI and ensemble learning models can provide more personalized and accurate career recommendations.

However, existing systems often:

• Lack personalization.
• Depend on limited datasets.
• Ignore behavioral and interest-based factors.
• Do not provide post-prediction career roadmaps.
• Lack interactive and user-friendly interfaces.

Methodology

The proposed system follows a structured workflow:

1. Data Collection
   • Career datasets containing skills, career domains, and job roles.
   • User questionnaire responses capturing interests and preferences.
   • Educational background data.
   • Job-role datasets linking skills to occupations.
2. Data Preprocessing
   • Missing value handling and imputation.
   • Skill standardization and encoding.
   • Feature normalization using Min-Max Scaling.
   • Outlier detection using Z-score and IQR methods.
3. Data Transformation
   • 80:20 train-test split with stratified sampling.
   • Label Encoding for career fields.
   • MultiLabelBinarizer for skill-set encoding.
4. Feature Selection
   • Correlation Analysis to remove redundant features.
   • Principal Component Analysis (PCA) for dimensionality reduction.
   • SHAP analysis to identify the most influential features.

Key Features Used

The model considers multiple types of user information:

• Technical Skills: Programming, data analysis, and domain-specific expertise.
• Soft Skills: Communication, teamwork, and problem-solving abilities.
• User Interests: Preferred fields such as IT, Business, or Design.
• Academic Performance: Educational background and subject strengths.
• Career Preferences: Desired roles and career aspirations.

Machine Learning Models

Stage 1: Random Forest

Used for career field prediction because it:

• Handles high-dimensional data effectively.
• Reduces overfitting through ensemble learning.
• Supports both categorical and numerical features.

Stage 2: LightGBM

Used for job-role recommendation because it:

• Offers high prediction accuracy.
• Trains efficiently on large datasets.
• Uses memory efficiently.

Hyperparameters such as number of trees, maximum depth, learning rate, and number of leaves were optimized to improve performance.

System Architecture

The platform consists of three layers:

1. Input Layer
   • Collects user survey responses and skill selections.
2. Processing Layer
   • Performs preprocessing, feature encoding, and machine learning inference.
3. Output Layer
   • Displays:
     • Predicted career field.
     • Top three job role recommendations.
     • Personalized career roadmap.

The system is supported by UML diagrams, including class, state, use-case, sequence, and block diagrams, to document system structure and workflow.

Implementation

The platform was developed using:

• Python 3.10
• Streamlit for the web interface
• Scikit-learn for machine learning
• LightGBM for job recommendation
• Pandas and NumPy for data processing
• Matplotlib, Seaborn, and Plotly for visualization

The implementation includes modules for:

• Data preparation.
• Feature encoding.
• Model training.
• Career prediction.
• Roadmap generation.

Testing and Validation

Seven test cases were conducted to verify:

• Data loading.
• Preprocessing.
• Feature extraction.
• Train-test splitting.
• User input encoding.
• Career prediction.
• Roadmap generation.

All tests passed successfully, confirming the correctness and reliability of the system.

Results and Discussion

The Smart Career Guidance Platform successfully provides three key outputs:

1. Career Field Prediction
   • Random Forest predicts the most suitable career domain.
2. Job Role Recommendation
   • LightGBM recommends the top three job roles aligned with the user's skills and interests.
3. Personalized Career Roadmap
   • Provides step-by-step learning paths, required topics, and certifications.

The two-stage prediction process improves recommendation relevance by first identifying a suitable career field and then narrowing job recommendations within that domain.

Conclusion

This paper has presented a Smart Career Guidance Platform leveraging a two-stage machine learning pipeline comprising Random Forest and LightGBM to deliver personalized career field predictions and job role recommendations. The system achieves approximately 90% accuracy for career field classification, demonstrating the practical effectiveness of ensemble machine learning in the career guidance domain. The platform addresses key limitations of existing systems by integrating multiple input modalities including technical skills, soft skills, behavioral preferences, and academic background; providing structured post-prediction career roadmaps with learning phases, topics, and certifications; and delivering all functionalities through an accessible Streamlit web application. The comprehensive testing protocol confirmed functional correctness and reliability across all seven test scenarios. The proposed system demonstrates that AI-driven, data-centric career guidance tools have significant potential to assist students and early-career professionals in making informed, evidence-based career decisions, reducing uncertainty, and improving alignment between individual competencies and industry requirements.

References

[1] A. O. Abisoye, I. O. Alabi, S. O. Ganiyu, B. O. Abisoye, and J. Omokore, “A web-based career guidance information system for pre-tertiary institution students in Nigeria,” Int. J. Sci. Res. Sci. Eng. Technol. (IJSRSET), Jun. 2015. [2] H. F. El-Sofany and M. S. A. El-Seoud, “The implementation of career and educational guidance system (CEGS) as a cloud service,” Int. J. Emerg. Technol. Learn. (iJET), vol. 15, no. 20, Oct. 2020. [3] A. J. García, A. Sneyd, A. Melro, and A. Ollagnier, “C3 IoC: A career guidance system for assessing student skills using machine learning and network visualization,” Int. J. Artif. Intell. Educ., Dec. 2022. [4] S. Gokarn, S. Taware, and R. Vartak, “Smart career guidance system using machine learning,” Int. J. Sci. Res. (IJSR), Jun. 2024. [5] R. Kanathur, R. Akhilesh, A. Soumya, and M. K. Huang, “Intelligent career guidance systems,” Int. J. Innov. Sci. Res. Technol., Aug. 2023. [6] A. Dahanke, N. Shinde, A. Dhagate, and H. Shaikh, “An intelligent career guidance system using machine learning,” Int. Res. J. Eng. Technol. (IRJET), vol. 9, no. 3, Mar. 2022. [7] S. Gunwant, J. Pande, and R. Bisht, “A systematic study of the literature on career guidance expert systems for students: Implications for ODL,” J. Learn. Dev., Nov. 2022. [8] J. L. Holland, Making Vocational Choices: A Theory of Vocational Personalities and Work Environments. Odessa, FL: Psychol. Assess. Resour., 1997. [9] D. E. Super, “A life-span, life-space approach to career development,” J. Vocat. Behav., 1980. [10] R. W. Lent, S. D. Brown, and G. Hackett, “Social cognitive career theory,” J. Vocat. Behav., 1994. [11] M. L. Savickas, “Career construction theory and practice,” in Career Development and Counseling, 2005. [12] P. Kaur and M. Sharma, “Career guidance system using machine learning techniques,” Int. J. Comput. Appl., 2021. [13] A. Singh and R. Kumar, “Career recommendation system using artificial intelligence,” Int. J. Eng. Res. Technol. (IJERT), 2020. [14] D. Patel and P. Shah, “A survey on career prediction using machine learning,” Int. J. Adv. Res. Comput. Sci., 2022. [15] S. Gupta and A. Verma, “AI-based career counseling system for students,” Int. J. Comput. Sci. Inf. Technol., 2023.

Copyright

Copyright © 2026 Dr. K. Subbarao, Dabbakuti John Victor, Murugula Shanmuk, Jidugu Bhargava Adithya, Bodipaga Siva Nagaraju. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.