Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AI-Driven Decision-Making System for Hiring Process

Vira Filatova, Andrii Zelenchuk, Dmytro Filatov

TL;DR

This work tackles the inefficiency of early-stage candidate validation by introducing a modular, multi-agent AI hiring assistant steered by an LLM. It integrates multimodal inputs (resumes, code tasks, video responses) with public-data verification, explicit risk penalties, and a human-in-the-loop interface to produce traceable, configurable candidate scores. The system demonstrates notable efficiency gains—1.70 hours per qualified candidate—while maintaining alignment with an experienced recruiter’s decisions and reducing screening costs, all within a transparent, configurable pipeline. The study highlights practical benefits for small, technically proficient companies and discusses ethical, privacy, and generalization considerations, outlining future directions for broader validation and deployment.

Abstract

Early-stage candidate validation is a major bottleneck in hiring, because recruiters must reconcile heterogeneous inputs (resumes, screening answers, code assignments, and limited public evidence). This paper presents an AI-driven, modular multi-agent hiring assistant that integrates (i) document and video preprocessing, (ii) structured candidate profile construction, (iii) public-data verification, (iv) technical/culture-fit scoring with explicit risk penalties, and (v) human-in-the-loop validation via an interactive interface. The pipeline is orchestrated by an LLM under strict constraints to reduce output variability and to generate traceable component-level rationales. Candidate ranking is computed by a configurable aggregation of technical fit, culture fit, and normalized risk penalties. The system is evaluated on 64 real applicants for a mid-level Python backend engineer role, using an experienced recruiter as the reference baseline and a second, less experienced recruiter for additional comparison. Alongside precision/recall, we propose an efficiency metric measuring expected time per qualified candidate. In this study, the system improves throughput and achieves 1.70 hours per qualified candidate versus 3.33 hours for the experienced recruiter, with substantially lower estimated screening cost, while preserving a human decision-maker as the final authority.

AI-Driven Decision-Making System for Hiring Process

TL;DR

This work tackles the inefficiency of early-stage candidate validation by introducing a modular, multi-agent AI hiring assistant steered by an LLM. It integrates multimodal inputs (resumes, code tasks, video responses) with public-data verification, explicit risk penalties, and a human-in-the-loop interface to produce traceable, configurable candidate scores. The system demonstrates notable efficiency gains—1.70 hours per qualified candidate—while maintaining alignment with an experienced recruiter’s decisions and reducing screening costs, all within a transparent, configurable pipeline. The study highlights practical benefits for small, technically proficient companies and discusses ethical, privacy, and generalization considerations, outlining future directions for broader validation and deployment.

Abstract

Early-stage candidate validation is a major bottleneck in hiring, because recruiters must reconcile heterogeneous inputs (resumes, screening answers, code assignments, and limited public evidence). This paper presents an AI-driven, modular multi-agent hiring assistant that integrates (i) document and video preprocessing, (ii) structured candidate profile construction, (iii) public-data verification, (iv) technical/culture-fit scoring with explicit risk penalties, and (v) human-in-the-loop validation via an interactive interface. The pipeline is orchestrated by an LLM under strict constraints to reduce output variability and to generate traceable component-level rationales. Candidate ranking is computed by a configurable aggregation of technical fit, culture fit, and normalized risk penalties. The system is evaluated on 64 real applicants for a mid-level Python backend engineer role, using an experienced recruiter as the reference baseline and a second, less experienced recruiter for additional comparison. Alongside precision/recall, we propose an efficiency metric measuring expected time per qualified candidate. In this study, the system improves throughput and achieves 1.70 hours per qualified candidate versus 3.33 hours for the experienced recruiter, with substantially lower estimated screening cost, while preserving a human decision-maker as the final authority.
Paper Structure (23 sections, 4 equations, 3 figures, 2 tables)

This paper contains 23 sections, 4 equations, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Communications module
  4. Resume evaluation
  5. Technical assignment evaluation
  6. Visual evaluation
  7. Answer evaluation
  8. Public data evaluation
  9. Culture fit evaluation
  10. Integrated multi-modal hiring assistants
  11. Methodology
  12. Preprocessing
  13. Initial Context
  14. Context Augmentation
  15. LLM Analysis
  16. ...and 8 more sections

Figures (3)

  • Figure 1: Schematic overview of the proposed modular multi-agent pipeline for candidate data ingestion, context construction, public-data verification, scoring, ranking, and validation.
  • Figure 2: Comparison of average review time per qualified candidate for expert recruiters, standard recruiters, and the proposed system.
  • Figure 3: Comparison of average review cost per qualified candidate for expert recruiters, standard recruiters, and the proposed system.