Operationalizing Fit in Recruitment: A Multi-KPI, Evidence-Based Matching Architecture

Angelo Leogrande, Mauro Di Molfetta (), Nicola Magaletti (), Valeria Notarnicola () and Stefano Mariani ()
Additional contact information
Nicola Magaletti: LUM - Università LUM Giuseppe Degennaro = University Giuseppe Degennaro
Valeria Notarnicola: LUM - Università LUM Giuseppe Degennaro = University Giuseppe Degennaro
Stefano Mariani: LUM - Università LUM Giuseppe Degennaro = University Giuseppe Degennaro

Working Papers from HAL

Abstract: The current study proposes a multi-KPI approach for the alignment of job offers with candidate profiles that combines semantic signals, skill coverage indicators, and behavioral/contextual dimensions into a single decision support architecture. The approach is based on a semantic-first paradigm that represents job offers and candidate profiles as text embeddings that can be compared using the cosine similarity measure to produce a scalable baseline ranking that remains robust across different languages and styles. The baseline approach can then be extended using additional KPIs that capture different dimensions of the candidate-job alignment: the Hard Skill Coverage Ratio (HSCR) and the Skill Gap Index (SGI) measure the satisfaction of hard skill requirements and the remaining skill gaps; the Hard Skill Proficiency Similarity (HSPS) measures semantic similarity in the hard skill domain; the Soft Skill Semantic Alignment (SSSA) and the Soft Skill Evidence Density (SSED) measure the semantic similarity of the candidate behavior and the quality of the corroborating evidence; the Cultural & Team Fit Score (CTFS) addresses the organizational fit; and other indicators cover the operational feasibility of the candidate-job match. The approach is implemented using Python and tested using a realistic testbed of 300 EURES job postings and 300 heterogeneous candidate profiles across different formats and languages. The results show that the semantic approach yields a stable baseline that can be used to produce coherent and well-ordered shortlists of candidate-job pairs while the additional KPIs improve the interpretability of the approach and the feasibility of constraint satisfaction. The analysis of the candidate-job scores and the separations of the rankings also reveals that there are clusters of technically equivalent candidates and that there are clear cases of dominance. Overall, the approach can support an incremental move toward multicriteria decision making while balancing scalability, transparency, and governance-by-design requirements.

Keywords: Job-candidate matching semantic embeddings multi-KPI framework skill coverage decision support systems; Job-candidate matching; semantic embeddings; multi-KPI framework; skill coverage; decision support systems (search for similar items in EconPapers)
Date: 2026-02-17
Note: View the original document on HAL open archive server: https://hal.science/hal-05514751v1
References: Add references at CitEc
Citations:

Downloads: (external link)
https://hal.science/hal-05514751v1/document (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:hal:wpaper:hal-05514751

Access Statistics for this paper

More papers in Working Papers from HAL
Bibliographic data for series maintained by CCSD ().