Cynor helps companies automate their cv screening process by uploading a job description and the corresponding CVs of all the applicants. After processing the data, Cynor will produce a ranked list of candidates and explain why the person received a certain rank.

Our main approach is to detect and match entities of importance to the CV screening task at hand. At first we try and detect 4 categories of information from the Job Description, namely:

Designation

Location

Requirements

Responsiblities (SBD)

Although, The system does not require all 4 categories to work but having all 4 improves accuracy. However, atleast requirements is necessary for it to work.

Then we proceed to extract entities or sentences from each category of information using Named Entity Recognition (NER) and Sentence Boundary Detection (SBD) techniques.

There are several ways in which a CV and JD map to one another. For our purposes, the following diagram shows how we think the categories in the JD map to the information in the CV and how we can use that to figure out the eligibility of that CV for that particular Job Description.

Link to project: https://github.com/cynor-codesec