E-CARE: Resume Screening, Candidate Matching Software

Application

Natural language processing (NLP) platform to facilitate more efficient, accurate, and non-biased identification of job candidates against open employment opportunities.

Key Benefits

  • Greatly reduces recruiter involvement.
  • Provides superior applicant-to-job matching accuracy.
  • Offers a plugin or API (Application Programming Interface) access point for easy integration into existing Applicant Tracking Systems.

Market Summary

An ongoing challenge for Human Resource (HR) recruiters is the screening involved to accurately match candidate qualifications to open job positions. Resumes and CV’s contain unstructured data that cannot be easily or intelligently processed for determining the relevancy of an applicant's experience in relation to a target employment position. The current practice of Boolean matching is limited in its ability to accommodate complex query syntax and the ability to accurately rank job candidates.

Technical Summary

The inventors propose E-CARE, a Natural Language Processing (NLP) platform to automate robust and non-biased identification of suitable candidates for open job descriptions. Empowered by the latest deep learning-based NLP techniques, E-CARE may reduce resume screening time by 15-fold while equaling (or bettering) the robustness of a candidate-to-position match performed by human recruiters. E-CARE may also provide reduced recruiter involvement in screening superior applicant-to-job matching accuracy over human supervised matching, and access points for web-APIs that can be easily integrated into Applicant Tracking Systems (ATSs).

E-CARE Workflow

Developmental Stage

  • The project has received $50,000 in Georgia Research Alliance funding to implement a “Minimal Viable Product” within Emory Healthcare Human Resources to better identify skilled clinical research coordinators and nurses.
  • To date, the inventors have adjudicated over 3,000 resumes using human subject matter experts to create a standard reference which the NLP model has been able to match. The intent is to process a larger number of resumes to improve the training accuracy of the model and deploy a working prototype by Spring of 2021.

Publication: Li, C. et. al. (2020). Competence-Level Prediction and Resume & Job Description Matching Using Context-Aware Transformer Models.

Patent Information

Tech ID: 20093
Published: 11/23/2020