Resume matching uses interpretable features distilled from large language models
Building Interpretable Feature Representations for Resume-Vacancy Matching by Distilling Production LLM Signals
Computation and Language
Summary
Finding good candidates for jobs often depends on showing clear reasons why a person matches a vacancy, not just a score. The authors developed a system that turns explanations from a large AI language model into simpler, understandable features that recruiters can see and use. They trained a smaller model to mimic the big model's outputs but run faster and on normal computers. Recruiters have tested this system and agreed with its feature explanations about 96% of the time.
What this means in practice
- •For recruitment technology teams: Integrate interpretable resume-vacancy matching features to assist recruiters in making hiring decisions with transparent evidence.
- •For human resources software developers: Deploy lightweight models that provide understandable matching criteria efficiently on standard CPUs without calling large language models online.
Authors
Ilya Chekin, Vyacheslav Malyugin, Vladimir Chirkov, Mikhail Yurushkin
Abstract
Matching candidates to vacancies is central to recruitment, and a recruiter needs to see why a candidate fits, not only a single opaque relevance score. We provide this evidence as named, interpretable matching dimensions recruiters can act on - eight in our current deployment. We propose a two-part approach. The first is an LLM-based labeler whose prompts and feature definitions were refined from recruiter feedback while it served as an earlier production matching stage. In the current architecture, it is used only for offline labeling and is not called on online requests. The second is a feature bi-encoder distilled from it: a LoRA-adapted embedding backbone with compact per-dimension heads that runs on CPU and serves all online requests. Both parts keep improving: prompts are revised as feedback arrives, and the bi-encoder is retrained on the updated labels. The model is trained on 168,772 labeled vacancy-resume pairs (17,921 vacancies and 180,030 resumes). Recruiters using the service can confirm or revise surfaced feature predictions. On 927 recruiter-recorded values from this selected production-feedback subset, the deployed student agrees with the recorded decisions in 888 cases (95.79%). This is operational, non-blinded agreement rather than an independent human evaluation.