Creating ML and AI systems in recruitment management system

ML and AI in the automation of work processes

Task:

System for recruiting and maintaining staff. Reduction of routine processes for selecting a candidate, structuring the processing of the resume, obtaining insights for "suitable" candidates.

Solution:

During the processing of the resume, we created our own algorithms, which took into account the specificity of multilingualism (Russian, Ukrainian, "mixed language"), as well as the relationship of textual phrases and their context for determining the skills and experience specific in a number of professions. We used ML technologies and databases on graphs to predict the candidate's compliance with a particular position and company. Hints on the behavioral parameters of HR managers of different companies (insights), autotune a number of routine processes for validated patterns of behavior (AI).

Project type:

All, Data Science & Business Analysis, Enterprise & Industry, Custom systems, Landing pages

Project status:

Ready

Languages:

RU, EN

Thematics:

Retail

Technologies:

Data Science, Symfony, React, jQuery

Link to website:

http://worxygen.com/

A huge list of functional, which was implemented:

A separate company account on the SaaS platform (link type: company.worxygen.net).

Adding users by invite (marketing goal: assessing customer inbound flows, recommendations and behavioral characteristics).
Structure of the company / departments / groups (formation of a tree-like type of nesting and the ability to change the structure).
Profiles of positions (both legal and factual).

What is the account for an employee within the SaaS platform:

Develop employees (using testing / polls) in detail, by the mass of direct and indirect parameters.
Evaluation of the slice of knowledge, attestation and tracking progress on them.

There is an opportunity to put KPI, evaluate the employee / manager and form forecasts of productivity / staff shortage in the "field".
Track patterns of employee behavior and make predictions.
To carry costs (employees can write down, the line manager can coordinate / reject).

Job requests were created. Separately the recruiter's calendar (applications for vacancies and deadlines). All the system receives notifications about new applications to the recruiter, new completed forms.

On the consolidated dashboard, you can view in a tabular form (or in a graphical format) generalized information on the status of open vacancies for the entire company. Find out who, how much is not in time or ahead of the plan.

The algorithm of the manual analyzer was specially developed to automate the scan of the CV by recruiters (with the ability to influence also in the manual mode).

Parsing resume from job websites (work.ua, rabota.ua, hh.ua, trud.ua and a number of others). A CV database has been created (storage, maintenance, vacancies, thematic tags). Download CV from a folder on your computer.

PlayMarket web forms (work on the brand - share the results of evaluating candidates with other companies and countries). Statistics of processed forms (with answers by anonymous users and employees).

Forms constructor was built with convenient creation of tests / questionnaires, mat. calculations. You can create the necessary fields yourself and drag-and-drop using drag-n-drop. Share the created web form and integration into another online service.