WORK - ATS

Applicant Tracking System

description

Background

ATS, It’s a human resources software that is basically a database for the job applicants. We have a product which makes recruiters and hiring managers life easier. It helps companies to communicate with a large group of applicants.

When a recruiter receives a lot of resumes, it’s not feasible for them to just go through all of them. At this stage, our product filters all the applications as per their requirements.

This allows businesses to collect information, organize prospects based on experience and skill set, and filter applicants.

Technologies Used

Backend:
PHP Laravel

Frontend:
HTML5, Css, Css3, SAAS, Bootstrap, Javascript, Ajax

Rest APIs:
Machine learning (Python), Django (Python)

Database:
MySQL

Our role
  • Front-End Engineering
  • Backend Engineering
  • Custom Design
  • Business Analysis
Category

Environmental

Timeline

3 Days

OVERVIEW

The Challenge

1. Parsing a resume and figuring out the skills, experience etc. from the resume is one of the major challenges. Skill keywords in the resume sometimes in the singular or in the pular form, therefore there may be a 50-50 chance of including the keyword from the resume.

2. Everyone has a different resume format. Considering this it's hard for a program to just read the resume the way you want. One can use experience heading to describe his or her experience in the resume, another one can use any other word to describe the same thing, and this possibility can be with any title in the resume.

3. Applicants applying for more than one position in the same organization are susceptible to a lower keyword percentage match, that means content writer, data scientist or any other position have different job description and skill set that can show less percentage match with the required skills.

4. Soft skills keywords are hard to identify from the resume like attention to details, good listener, adaptive learner etc.

Solution

1. Created a csv file having all the possible skills, then trained a model to read all the skills and match those skills from the resume. After applying the trained model to our software, we are at the stage of figuring out around 95% of the applicant skills accurately as in the resume.

2. For reading all the required titles from the resume, we have defined the constants and created a dataset that stores the titles and the possible values, then a model is trained and implemented over the software. Final parsed output from the resume is Name, Email, Mobile Number, skills, experience, education.

how we work

Our Strategy

Create

Our designers and architects work with you to define every feature, screen, and user flow.

Sprint

Receive product builds every two-weeks as we add features.

Learn

We analyze user feedback to help you prioritize new features.

preview

Product Overview

Applicant Tracking System

Quick Service

We complete Phase 1 ESAs faster than other firms without sacrificing quality performance.

Quick Service

We complete Phase 1 ESAs faster than other firms without sacrificing quality performance.

Quick Service

We complete Phase 1 ESAs faster than other firms without sacrificing quality performance.