Αρχιτεκτονική και εξατομίκευση λογισμικού (backend)

Μπαρμπούνη, Ειρήνη

Architecture and personalization for backend applications

Master Thesis

Author

Μπαρμπούνη, Ειρήνη

Date

2026-01

Abstract

This Master’s thesis focuses on the design, analysis, and implementation of a recommendation system that supports personalized candidate-job matching through a rule-based evaluation engine. The system is positioned within the domain of intelligent human resource management, addressing the growing need for automation and personalization in recruitment processes. Modern labor markets are increasingly dynamic and data-driven, requiring intelligent software capable of evaluating, filtering, and adapting recommendations for both candidates and recruiters. The application has been developed following the principles of the Hexagonal Architecture (Ports & Adapters), aiming to achieve extensibility, maintainability, and technology independence. This architecture ensures a clear separation between business logic and external systems, allowing for the replacement of technologies (such as databases, APIs, or recommendation engines) without affecting the system’s core functionality. This results in a loosely coupled, highly cohesive structure that supports long-term scalability and flexibility. At the heart of the system lies the recommendation engine, which calculates compatibility scores between candidates and job positions based on hard and soft skills collected through structured questionnaires. The resulting scoring mechanism enables a quantitative assessment of suitability and produces personalized job recommendations for each candidate. Recruiters, in turn, can automatically receive ranked candidate lists per job position, facilitating faster and more objective decision-making in the hiring process. To ensure adaptability, the system employs the Strategy Pattern, enabling the dynamic selection of different recommendation algorithms—for instance, switching between rule-based and AI-driven methods—without modifying the core architecture. This approach supports continuous experimentation and future integration of machine learning or feedback-based learning mechanisms, enhancing the personalization capabilities of the system. The backend has been developed using Spring Boot, structured around domain services and defined ports, exposing functionalities via RESTful APIs. These APIs can be consumed by any user interface, whether web or mobile, offering a flexible and interoperable foundation. The system design process included requirements analysis, UML modeling, and the development of the complete backend flow, accompanied by data flow diagrams, code samples, and user interface snapshots. 5 The thesis concludes with a discussion on potential future extensions, such as integrating AI and ML-based recommendation models, implementing adaptive learning from user feedback, and supporting A/B testing to evaluate multiple recommendation strategies. These enhancements would allow the system to evolve into a more intelligent and adaptive platform for personalized recruitment. Ultimately, this research demonstrates how modern software architectural principles and design patterns can be effectively combined to create systems that are modular, scalable, and future-ready, paving the way toward smarter and more personalized digital solutions in the field of employment and human capital management.

Postgraduate Studies Programme

Προηγμένα Συστήματα Πληροφορικής

Department

Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Πληροφορικής

Number of pages

Language

Greek

URI

https://dione.lib.unipi.gr/xmlui/handle/unipi/19003

Collections

Τμήμα Πληροφορικής

Show full item record

Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα