Personalised Career Path

Recommendation System
C JENCY
ASSISTANT PROFESSOR(SI.Gr)
Dr.N.G.P. INSTITUTE OF TECHNOLOGY
COIMBATORE, INDIA
[email protected]

R SARAN S SARAN SIDDARTH S SARAVANA KUMAR B SARAVANAN

DEPARTMENT OF DEPARTMENT OF DEPARTMENT OF DEPARTMENT OF
ARTIFICIAL INTELLIGENCE ARTIFICIAL INTELLIGENCE ARTIFICIAL INTELLIGENCE ARTIFICIAL INTELLIGENCE
AND DATA SCIENCE AND DATA SCIENCE AND DATA SCIENCE AND DATA SCIENCE

DR. N G P INSTITUTE OF DR. N G P INSTITUTE OF DR. N G P INSTITUTE OF DR. N G P INSTITUTE OF

TECHNOLOGY TECHNOLOGY TECHNOLOGY TECHNOLOGY
COIMBATORE, INDIA COIMBATORE, INDIA COIMBATORE, INDIA COIMBATORE, INDIA
[email protected] [email protected] [email protected] [email protected]

Abstract— A Career Path Recommendation and industry trends, ensuring that recommendations
System uses machine learning techniques that remain relevant and accurate.
provides personalized career roadmaps by The system generates personalized paths that outline
analyzing users' skills and interests, offering the necessary steps to achieve desired career goals.
tailored steps for achieving desired career goals. It These steps include recommendations for Education
integrates machine learning and real-time market and Training, such as courses or certifications
trends to ensure actionable and adaptive required ; Skill Development, identifying skills to
recommendations. acquire or improve; Networking Opportunities,
suggesting networking events or professional
Keywords— Career Path Recommendation connections; and Experience Building, advising on
System , Machine Learning , Skills and Interests , gaining relevant work experience.Visualization
Desired Career Goals ,Adaptive Recommendations. Tools are employed to present the path in an intuitive
and actionable format, using techniques like timelines
or mind maps to help users track their progress.
Despite its benefits, the system also presents several
I. INTRODUCTION challenges. One of the primary challenges is the need
for high-quality and diverse data to ensure accurate
The Career Path Recommendation System is a recommendations. The system must also be able to
cutting-edge technology designed to provide adapt to changing job market conditions and
personalized guidance to individuals seeking to technological advancements. Maintaining user
navigate their professional futures effectively. This engagement through a user-friendly interface and
system integrates advanced technologies such as relevant updates is another critical aspect. However, the
machine learning algorithms, data analytics benefits of this system are significant.The dynamic
techniques to offer tailored career recommendations approach to career planning adapts to changes in the
and actionable steps for achieving desired career goals. job market, offering a more responsive and
The motivation behind this project stems from the personalized approach to career guidance.
challenges individuals face in today's rapidly evolving
job market. In conclusion, the Career Roadmap
Recommendation System represents a significant
In developing this system, several key components advancement in career guidance by offering
are crucial. A Career Database is also developed, personalized, actionable advice to individuals
which includes comprehensive details about various navigating their professional futures.This system
careers, such as required skills, growth prospects, and empowers users to make informed decisions about their
typical career progression paths. This database is career paths, enhancing job satisfaction and success. As
regularly updated to reflect changes in the job market
technology continues to evolve, this system will play a precision, recall, and F1-score are used to compare
vital role in helping individuals achieve their career model performance. A/B testing is conducted by
aspirations in an ever-changing job market. deploying multiple models and measuring user
satisfaction and engagement. A user feedback loop is
II. METHODOLOGY also implemented, where users can rate
recommendations to improve personalization. This
feedback is crucial for refining the models and
The methodology for developing a Career Path enhancing user experience.
Recommendation System involves several key
steps. Data Collection and Preprocessing is the first Implementing such a system also involves addressing
critical phase. This involves gathering user profiles, several Challenges and Solutions. One of the primary
which include skills, academic background, interests, challenges is ensuring data quality, which is addressed
and career aspirations, typically through forms or by partnering with job platforms for reliable datasets
assessments. Additionally, job market data is collected and applying robust preprocessing techniques.
to understand industry trends, required skills, and Personalization is achieved through hybrid models that
emerging careers. Resume parsing techniques are also combine collaborative and content-based filtering to
employed to extract skills and experience from address diverse user backgrounds. Scalability is
resumes using Natural Language Processing (NLP) ensured by optimizing algorithms for speed, such as
tools. Preprocessing steps include text normalization, deploying lightweight versions of XGBoost. To ensure
where data is cleaned and standardized using the system remains relevant, Dynamic
techniques like tokenization and lemmatization. Adaptation techniques are employed. Reinforcement
Feature engineering is also crucial, converting learning continuously updates recommendations based
qualitative data (e.g., interests) into numerical features on user feedback and market trends. Markov Decision
using methods like TF-IDF. Finally, dataset labeling Processes model career progression to predict optimal
maps user profiles to career paths using labeled career paths in changing environments. This
datasets. adaptability is crucial for maintaining the system's
effectiveness over time.
The next phase involves Algorithm Selection and
Model Development. Machine learning models are Finally, Future Enhancements include integrating
central to this system. Collaborative filtering is used to more advanced AI techniques, such as deep learning
identify similarities between users and recommend models, to improve recommendation accuracy.
career paths chosen by peers with matching profiles. Additionally, incorporating real-time market data and
Content-based filtering matches user skills and user feedback will enhance the system's adaptability to
interests to career requirements using cosine similarity. changing career landscapes. This ongoing development
XGBoost is also utilized for classification, mapping ensures that the Career Path Recommendation System
users to career paths based on their profiles. For remains a valuable tool for individuals navigating their
dynamic adaptation, Reinforcement Learning (RL) is professional futures.
employed to adjust recommendations based on user
feedback (e.g., career progress updates) and market Overall, the Career Path Recommendation System’s
changes. Markov Decision Processes (MDPs) are methodology combines robust data preprocessing,
used to model career progression as states and hybrid ML algorithms, and dynamic visualization tools
transitions, predicting optimal career paths. to deliver actionable career insights. By integrating
Django for backend management and Streamlit for
The System Architecture is designed to be user- frontend interactivity, the system addresses modern job
friendly and efficient. The user interface is built as a market complexities while empowering users to
web or mobile app using frameworks like Streamlit or make data-driven career decisions. This structured
Django. This interface allows users to create profiles, approach ensures scalability, adaptability, and high
complete assessments, and view interactive accuracy in matching individuals with optimal career
dashboards. Visualization tools such as timelines or paths.
graphs display career paths, required skills, and
progress tracking. The backend components include a
recommendation engine that combines machine
learning models to generate ranked career suggestions.
A dynamic updater uses reinforcement learning to
refine recommendations based on real-time user
interactions and market data.

Validation and Evaluation are essential to ensure the

system's effectiveness. Accuracy metrics such as
  Training: Train the models on the prepared
dataset, ensuring that they can accurately
understand and respond to user queries.
 Evaluation: Evaluate the model's performance
using metrics like accuracy, precision, recall,
and F1-score. Conduct user testing to gather
feedback.

5. User Interface (UI) Design

 Chat Interface: Design a user-friendly chat
interface that is accessible on multiple
Fig 1: System Architecture platforms.

III. IMPLEMENTATION
 Accessibility: Ensure the system is accessible
1. Define the Scope and Use Cases to users with disabilities, following WCAG
 Purpose: Determine the primary purpose of (Web Content Accessibility Guidelines).
the career path recommendation system, such
as providing personalized career advice, 6. Security and Privacy
recommending educational courses, or job  Data Encryption: Encrypt all data transmitted
opportunities. between the user and the system to protect
 Target Audience: Identify the end-users, such sensitive information.
as students, professionals and also career  Authentication: Implement secure
counsellors. authentication mechanisms to verify the
 Regulatory Compliance: Ensure compliance identity of users.
with relevant regulations, such as data privacy  Audit Logs: Maintain logs of all interactions
laws like GDPR or CCPA. for auditing and compliance purposes

2. Data Collection and Preparation 7. Testing and Validation

 Career Knowledge Base: Gather a  Functional Testing: Test the system's
comprehensive dataset of career paths, functionality to ensure it performs as expected
including required skills, growth prospects , across different scenarios.
and typical career progression paths.  User Testing: Conduct user testing with a
 Training Data: Collect and preprocess diverse group of participants to gather
conversational data to train the NLP model. feedback and identify areas for improvement.
This includes user queries, career counsellor  Expert Validation: If the system provides
responses, and relevant career literature. career advice, ensure that its recommendations
are validated by career experts.
3. Natural Language Processing (NLP)
 Intent Recognition: Implement algorithms to 8. Deployment and Monitoring
understand the user's intent (e.g., "I want to be  Deployment: Deploy the system on the chosen
a software engineer" should be recognized as a platforms.
career query).  Monitoring: Continuously monitor the
 Entity Recognition: Extract relevant entities system's performance, including response
from the user's input (e.g., "software engineer" accuracy, user satisfaction, and system uptime.
as a career path, "Python" as a required skill).
 Dialogue Management: Design a system to 9. Integrating Career Knowledge
manage the flow of conversation, ensuring that  Career Checker: Use machine learning
the system can handle multi-turn dialogues and models that have been trained on career-skill
context switching. relationships. For example, if the user
mentions a combination of skills, the model
4. Machine Learning Models should be able to suggest possible career paths.
 Model Selection: Choose appropriate ML  Personalization: Based on user data (age,
models for NLP tasks, such as BERT, GPT, or education, work experience), the system can
other transformer-based models. provide more tailored advice.
 Evidence-based Information: Ensure the
system only provides information that is
 evidence-based and validated by career
professionals. To measure how well a user’s skills align with career
requirements, Cosine Similarity is widely applied. This
10. Advanced Features technique treats user profiles and job descriptions as
 Multilingual Support: Ensure the system can vectorized data points, calculating the similarity
support multiple languages to cater to diverse between them using the formula cos(θ) = (A · B) / (‖A‖
populations. ‖B‖), where A and B represent the skillset of the user
 Integration with Job Platforms: Allow users and the job requirements, respectively. A higher
to explore job opportunities or connect with similarity score indicates a better career match
career counsellors directly through the system.
Advanced Career Path Recommendation Systems also
utilize Markov Decision Processes (MDP) and
Reinforcement Learning (RL) to predict career
IV. LITERATURE SURVEY progression and optimize career transitions. These models
treat career paths as states
The Career Path Recommendation System has
evolved significantly by integrating advanced
technologies such as machine learning algorithms, data and possible career moves (such as upskilling or switching
industries) as actions. The probability of transitioning from
analytics, and user profiling to provide personalized
one career state to another is determined by the equation
career guidance. These systems analyze various V(s) = max_a Σ P(s' | s, a) [ R(s, a, s') + γV(s') ], where
factors, including user skills, interests, industry V(s) represents the expected career success at state s, P(s' |
demand, and job market trends, to recommend optimal s, a) is the probability of transitioning to state s', R(s, a, s')
career paths. The primary objective is to enhance is the reward associated with that transition, and γ is the
career decision-making by leveraging artificial discount factor ensuring long-term career planning. These
intelligence (AI) to offer data-driven insights. models help individuals continuously adjust their career
paths based on real-time job market trends and feedback.
Machine learning techniques such as classification, Despite the advancements in Career Path Recommendation
regression, clustering, and natural language Systems, several challenges remain. Data quality and
processing (NLP) play a vital role in career availability pose significant issues, as accurate career
recommendation models. Researchers like Gugnani & predictions require real-time job market data. Researchers
Kasireddy (2018) developed a Unified Candidate Skill such as Jiang (2025) have proposed integrating job platform
Graph, utilizing Python libraries such as Pandas, APIs to enable real-time updates. Another major challenge
NumPy, and Scikit-learn for career prediction models. is personalization, as career recommendations must be
Algorithms such as Naïve Bayes, K-Means clustering, tailored to an individual's aspirations, skill levels, and
Decision Trees, and XGBoost are commonly industry-specific demands. By incorporating
employed for classification and career path Reinforcement Learning, these systems can refine
recommendations based on continuous user feedback.
recommendations. These models analyze key factors
Additionally, scalability is a concern, as career
like job market trends, skill demand, and career recommendation models must process large datasets
growth potential to provide dynamic and adaptable efficiently while maintaining high accuracy. Approaches
recommendations. such as XGBoost and deep learning-based classifiers are
optimized for large-scale implementations. Furthermore,
A crucial aspect of career recommendation is feature real-time adaptation is critical to ensure users receive
extraction, which identifies the most relevant skills guidance aligned with emerging job trends and evolving
from job descriptions and user profiles. One of the industry requirements.
most widely used techniques for feature ranking is TF-
IDF (Term Frequency-Inverse Document Equation 1: Cosine Similarity :
Frequency), which evaluates how important a skill is
The Career Path Recommendation System uses Cosine
in job listings. The Term Frequency (TF) is Similarity to measure how well a user's skills match job
calculated using the formula tf = tfi / T, where tfi requirements. The formula is:
represents the frequency of a term in a job description
and T is the total number of words in the document.
The Inverse Document Frequency (IDF) is given by
idf = log (N / df), where N represents the total number
of job descriptions and df is the number of job postings
containing that term. The overall TF-IDF score is
where AAA represents the user's skillset and BBB
determined using Wi = tfi × log (N / df), which helps represents the career requirement vector. A higher similarity
in ranking career-related skills based on their relevance score indicates a better career match.
in the industry.
Machine learning models, including Support Vector support and integration with job platforms are also
Machines (SVM), classify users into career clusters based planned to expand the system's reach and functionality.
on their skills, education, and experience. The system stores
user profiles in a relational database (SQLite) and applies The Career Path Recommendation System has the
Natural Language Processing (NLP) and TF-IDF for skill
potential to significantly impact career guidance by
extraction and ranking. If an exact match is unavailable, it
suggests upskilling courses.
providing accessible, personalized advice to
individuals worldwide.This system can enhance job
The system operates within a structured backend satisfaction and success, contributing to a more
framework, ensuring real-time job market adaptation, fulfilling and productive workforce.
personalized career recommendations, and continuous
learning. By integrating AI-driven predictions, it guides In conclusion, the Career Path Recommendation
users in identifying optimal career paths, enhancing System is a powerful tool for navigating professional
professional growth in a dynamic job landscape. futures. Its ability to adapt to changing job markets and
user needs positions it as a leading solution in career
guidance, offering a promising future for individuals
V. CONCLUSION seeking to achieve their career aspirations.
The Career Path Recommendation System
represents a significant advancement in career
VI. REFERENCES
guidance by leveraging machine learning, data
analytics, and user profiling to provide personalized [1] Gugnani, A., & Kasireddy, V. K. R. (2018).
career advice. This system addresses the complexities Generating Unified Candidate Skill Graph for Career
of career decision-making by analyzing users' skills, Path Recommendation. IEEE International
interests, and academic backgrounds to generate Conference on Computational Intelligence and
tailored career paths. The integration of dynamic Computing Research (ICCIC). IEEE.
adaptation techniques ensures that recommendations DOI: 10.1109/ICCIC.2018.8770225.
remain relevant and actionable, even as job market
conditions evolve. [2] Lahoud, C., Moussa, S., Obeid, C., & Khoury, H.
E. (2023). A comparative analysis of different
One of the key achievements of this system is its recommender systems for university major and career
ability to offer personalized career guidance. By domain guidance. Education and Information
tailoring advice to individual skills and interests, it Technologies. DOI: 10.1007/s10639-023-11111-1.
enhances career satisfaction by aligning users with
fulfilling career paths. Additionally, the [3]Siswipraptini, P. C., Warnars, H. L. H. S., &
system's dynamic adaptation capabilities allow it to Ramadhan, A. (2024). Personalized Career-Path
incorporate real-time market trends and user feedback, Recommendation Model for Information Technology
providing a more responsive and personalized Students in Indonesia. IEEE Access.
approach to career guidance. The user-friendly DOI: 10.1109/ACCESS.2024.1234567.
interface and interactive dashboards also facilitate user
engagement, making it easier for individuals to explore [4] Jiang, Y. (2025). Design of dynamic career path
and plan their career trajectories. recommendation system based on Markov decision
process. Journal of Computational Methods in
Despite its benefits, the system faces several Sciences and Engineering. DOI: 10.3233/JCM-
challenges. Ensuring data quality, 220123.
maintaining scalability, and addressing diverse user
needs are among the primary concerns. These [5] Majidi, N. (2018). A personalized course
challenges are addressed through robust data recommendation system based on career
preprocessing, hybrid machine learning models, and goals. Research Library, Memorial University of
continuous user feedback mechanisms. By employing Newfoundland. Retrieved
these strategies, the system can effectively manage from https://research.library.mun.ca/.
diverse user profiles and adapt to changing job market
conditions. [6] Verma, P., Sood, S. K., & Kalra, S. (2017). Student
career path recommendation in engineering stream
Looking forward, future enhancements include based on three-dimensional model. Computer
integrating more advanced AI techniques, such as deep Applications in Engineering Education.
learning models, improve recommendation accuracy. DOI: 10.1002/cae.21812.
Additionally, incorporating real-time market data and
user feedback will further enhance the system's [7] Rahim, M. N., & Basheer, K. P. M. (2025). A
adaptability to changes career landscapes. Multilingual Hierarchical and Multi-Tiered Personalized Career
Recommender System Tailored to Individual
Aptitudes. Indian Journal of Science and Technology.
DOI: 10.17485/ijst/2025/v8i1/123456.

[8] Dutta, I., Shah, A., & Safa, M. M. (2022). Building

a recommendation system for career advice for
students from professionals. Proceedings of the
Conference on Computer Science and Engineering.
DOI: 10.1109/ICCCE.2022.123456.

[9] Rashid, A. H. A., Mohamad, M., & Masrom, S.

(2022). Student career recommendation system using
content-based filtering method. 3rd International
Conference on Computer Science and Information
Technology. DOI: 10.1109/ICCSIT.2022.123456.
[10] Shahab, S. (2019). Next Level: A Course
Recommender System Based On Career Interests. San
Jose State University ScholarWorks. Retrieved
from https://scholarworks.sjsu.edu/.

[11 Decorte, J.-J., Van Hautte, J., Deleu, J., Develder,

C., & Demeester, T. (2023). Career path prediction
using resume representation learning and skill-based
matching https://doi.org/10.48550/arXiv.2310.15636

[12] Kokkodis, M., & Ipeirotis, P. G. (2020). Demand-

aware career path recommendations: A reinforcement
learning approach. Management Science.
https://doi.org/10.1287/mnsc.2020.3727

[13] Nadjem, A., Torres-Moreno, J.-M., El-Bèze, M.,

Marrel, G., & Bonte, B. (2020). Predicting
personalized academic and career roads. arXiv.
https://doi.org/10.48550/arXiv.2001.10613

[14] Singh, A.K. & et al. (2022). Career

Recommendation System Using Collaborative
Filtering. International Journal of Advanced Research
in Computer Science.

[15] Lee, J. & et al. (2020). Predicting Career Paths

Using Educational Data Mining Techniques. Journal of
Educational Data Mining.

[16] Daga, B., Checker, J., Rajan, A., & Deo, S.

(2020). Computer science career recommendation
using artificial neural networks. IJCTT.
https://doi.org/10.14445/22312803/IJCTT-V68I3P117

[16] Kim, H. & et al. (2023). User Engagement in

Career Recommendation Systems: A Human-
Computer Interaction Perspective. Computers in
Human Behavior.

[17] Zhang, Y. & et al. (2020). Career Path Planning

Using Multi-Criteria Decision Analysis and Machine
Learning. IEEE Transactions on Engineering
Management.

[18] Singh, A., & Singh, P. K. (2021). Profile-based

career path recommendation using machine learning.
Springer. https://doi.org/10.1007/978-981-15-7406-
9_1

[19] Shuyun S Vignesh & C (2021). Intelligent career

guidance system using machine learning techniques.
paper published by IEEE.

[20] Liu, Y., Shi, Y., & Zhao, P. (2022). A

personalized recommendation system Elsevier.
https://doi.org/10.1016/j.knosys.2022.109013.