An AI-Powered SaaS product optimizes recruitment cycles for hiring teams worldwide

One main question that client needed to think was whether they should buy the AI application for HR or they should get the AI applications build for their own usage. Since client was not having deep experience in AI, hence taking into consideration Flexsin strong competency in building similar solution, Flexsin was asked to build the AI application from scratch.

Solution:

• Client main objective was to create a candidate enhancing experience that starts involving job seekers from the first interaction itself, while in parallel developing understanding of their suitability for roles that match their competency.
• Previously, jobseekers would meet for the first time at the job interview. By leveraging AI, jobseekers can now have real-time interaction via a chatbot, leading to more customized application process for job seekers. Job seekers with good amount of information of vacant positions in turn leads to a stronger fit of job applicants for roles.
• Videos got integrated into the process to give a much more information on hand to job seekers what it is like to work at the organization. Candidates were able to record video interviews from application's UI or their browser of choice and submit them for review
• High-level data engineering and data science capabilities were used for designing, developing, and deploying an AI module. This intelligent module had built-in capabilities of processing millions of CVs and scoring them according to the data available.
• Module was programed for comparing it with the job description the talent manager was looking for. Moreover, as most of the CVs may not be up to date, the application was designed to fetch the latest information about a specific candidate by connecting with a host of third-party apps. These third-party applications were the ones that collected the latest information and kept feeding the database.
• Managers can review more accurate job matches, job seekers can explore job openings relevant to their skills, and recruiters can find qualified applicants without spending time on manual research and job-to-applicant matching. Talent manager were able to schedule a video interview with any candidate and quickly view the social media profiles for every candidate

TECHNOLOGIES:
Node.js | Python| MongoDB |Express.JS |jQuery |Angular | Stripe Payment Gateway

Front-end
Flexsin's team designed a user interface that had the latest component of Angular in it. The front-end developers kept in mind the importance of building a simple and responsive UI.

Backend
Harnessing the power of Node and MongoDB, the backend on the website was fast and highly available for serving the user request as optimally as possible. Using the Shading concept of MongoDB for making sure the database queries were processed fast. Node clustering was leveraged for making the application highly available with zero downtime

Development Process
Below steps were taken for the development of AI-based application

We started with defining the use case and discussing it with business stakeholders on MVP. Then we checked the availability of data, carried out basic data exploration, defined the model-building methodology, established the model-validation methodology, and performed automation and production rollouts.

• This was a four-month project
• Weekly discussions were held between Flexsin team of data enginers, data scientists & talent acquisition team from client side.
• There were two in-person visits to client site for the objective of extensive process review and security considerations.
• Features were delivered sprint wise as per scrum methodology.

Since Flexsin also had a maintenance contract and hence discussion with the management on the continuous monitoring was done as once AI application has been published and deployed for use, it must be continuously monitored as, by understanding its validity. Proper training was organized from time to time to remain mindful of what process must be followed to ensure the model is kept up to date.

Data Provided to Flexsin:

• 20,000 job seekers resume in raw format
• 30 job descriptions from production department
• 10,000+ historical hiring decision points