Shine.com is Simplifying the Recruitment Process with AI and ML.

With 80lakh active candidates trying to find jobs and over 3lakh employment postings, Shine.com is India’s second-largest job portal.
“When recruiters post jobs on the portal to look for ideal applicants,
there has to be a right match of candidates on the bases of ability sets,” describes Madhukar Kumar, Chief Analytics Officer at Shine.com.
Showing results of applicants with comparable skill sets to the recruiter is exactly what Shine calls candidate-to-candidate matching.  In exactly the same fashion, there are 3 additional matching types: candidate-to-job, job-to-candidate, and job-to-job. With the proper algorithms at the back-end, the company finds relevant and suitable matches for candidates and recruiters both. Joining Shine 4 months back, Kumar has become positively working towards creating Shine an AI-first business,  and also bringing a data-driven culture in the organization.
“Earlier, some choices could have been taken based on gut feeling but right now we’re following a data-driven approach. Everything that people do right now, data is at the center of it,” states Kumar.
Kumar has additionally enacted an analytics roadmap for your business to resolve company issues using deep learning, ML, and AI.
To build a data science team with 6 data scientists and plans to grow it in the near future.
Based on Kumar, algorithms had been used earlier also but with the new team, the focus has grown on AI and also ML to improve search accuracy on the portal. Candidates in the organization that aspire to be data scientists have been identified.  With an aim to teach such engineers in data science, Kumar also conducts sessions with the organization during a weekly schedule.
The business plans to have many more than 17 data scientists until next year.
Kumar’s data science staff is enhancing the algorithms to improve the procedure of matching applicants with recruiters.
After posting a project, several recruiters get up to 1000 uses of candidates, making the filtering of relevant applications difficult and time-consuming. The business is focusing on ARS (Application Relevance Score) in which every applicant is going to get a score depending on the skill set necessary for the job published by the recruiter. This will lead to just the relevant applications ending in front of the recruiter, and
minimize the hiring cycle as well.
At current, the ARS is restricted to the IT business, however, there are blueprints to scale it to various other verticals also. Mapping Job titles (JT) with abilities is yet another area Kumar is working on. Kumar describes, If a candidate queries for data scientist jobs, there might be various other job titles related to it — for instance, associate or maybe guide data scientist, business analyst or perhaps data analyst.
A recruiter might possibly post a diverse job title but the abilities needed for those titles are far similar.
The task portal also shoots Job Alert Mailers (JAM) to applicants and that notifies them about new openings and outcomes in a higher selection of employment apps. Based on Kumar, making changes in matching algorithms has resulted in a meteoric increase in the number of users through JAM.