Banner
Banner10
What we did:

A Complete Web Portal

Technology and Performance Upgrade

Technologies Used :
Java, SpringCloud, Spring Data JPA, Angular4, Spring REST Services, Microservice Architecture

About Shortlist
Shortlist is a web portal that combines technology and human screening to find outstanding talent for businesses. It provides organizations with top qualified candidates based on the skills and attributes they demonstrate via in-depth performance analysis.
With world class resources at their disposal, Shortlist discovers, evaluates, informs and places talent across variety of industries. Candidates take online assessments to highlight their professional potential and based on that, they get matched with the right opportunity.

Challenges in Version 1

The previous version was mainly focused on testing the idea and it’s technical feasibility. But considering the scope and growth of the business, it was clear that we needed a major upgrade to the portal. The MVP needed a fix in various departments including Frontend and backend technologies, UI/UX, Performance and overall business model.
So we decided to map out ‘Version 2’ which will be more focused on achieving strict targets in terms of number of new sign ups and revenue generation. To achieve this, we did some major upgrades to the portal which resulted in achieving the targets from the first month itself, let’s decipher the magic step by step.

Major Upgrades Done

Technical Upgrades

We did major technological upgrades to the portal including replacing old protocols with the new ones and adding latest tech features.

UI Upgrades

We revamped the corporate website in major ways to make it more appealing to users with superior ‘User Experience’.

Assessment Upgrades

The screening process was fully automated while eliminating the manual work. Candidates get real time score and report.

Performance Upgrades

Greatly improved page load time from 200 candidates/3 seconds to 1000 candidates/second.

The Transformation Process

In Version1, we were relying on Java, Spring, Hibernate, Angular1 and Apache CXF. While in Version2, we replaced Angular1 by the recently introduced Angular4, added Spring Cloud to Spring technology, integrated Spring Data JPA and used Spring REST Services. Angular4 empowered developers to make the website better in terms of attainment and its response. Apart from that, we reformed the Login and Sign up screens. The changes are done in Login, Sign up and Reset Password screen made it easier for new users to get onboard. We also added the provision to login with Google account and integrated our website with Google recaptcha to protect it from spam and abuse.
We also took an important decision to move from Monolithic design pattern to a Microservice Architecture. The idea was to divide the single application into multiple interconnected applications.  Why? Well, a microservice architecture design pattern has many advantages over a monolithic design pattern including Scalability, Availability, Faster and more continuous Service deployment, Technology diversity, much better performance and above all, cost effectiveness.
Version 1 Version 2
Java, Spring, Hibernet, Angular1, Apache CXF, Monolithic Design Java, Spring+SpringCloud, Spring Data JPA, Angular4, Spring REST Services, Microservice Architecture

Upgrades in Screening Process

We also launched a dedicated operations portal to manage candidate screening life-cycle. We defined ‘15 Days’ as a duration to close the complete screening process. Candidate will receive an email along with the progress report and feedback at each stage including whether their application has been accepted or rejected, whether they are qualified or disqualified for further screening and their final selection.
The assessment will also be evaluated in real time and the result will be communicated immediately along with the comment. For example, ‘You couldn’t move ahead because you are overqualified for this position’.
If a candidate is qualified, then he/she is automatically redirected to the next assessment.
This automation resulted in huge time, cost and resource savings for the company. Not just that but the overall user experience throughout the screening process has also been improved by great extent.
Upon completion of the application process, candidate will be prompted to complete the personal profile and take assessments

Assessments

Previously, all assessments used to be carried out internally. This task is now assigned to a company that expertises in online assessments named ‘Interviewed’. The move helped Shortlist to focus on foundational activities and delegate the manual tasks to the specialised organisation, also freeing valuable time and resources.
Psychometric  Tests used to be conducted on third party software portal named ‘Mettl’. The decision was made to develop a fully functional automated tool similar to ‘Mettl’ and integrate it with Shortlist for Psychometric assessments. This helped to establish more control over the tests and make frequent updates from time to time.
New Assessment

Performance Upgrades

Version1 was lagging behind the competition when it comes to ‘Page Load Time’ and Loading time of candidates count. Previously it used to take 3-4 seconds to load a page with 200-300 candidate list. So we worked on this by streamlining the processes and replacing unwarranted ones with those which will help boost the loading speed and server communication. With the updations done, now it only takes 1-2 seconds to load the list of 1000 candidates. And the result is not just limited to candidate names but it also includes all the major fields and filters.

How we Helped the Platform become Scalable  using AWS Infrastructure

The idea was to develop a scalable and robust platform. Infrastructure was an important aspect to support new users and to help autoscale. We used Amazon EC2 container service to provide highly scalable, fast and easy to run web portal. Our process involved continuous evolution and improvement of the product. We focused on enhancing development cycle and fast deployment frequency
How challenges were dealt while implementing DevOps
  • Tweaking AWS infrastructure
The challenge ahead of us was to optimize AWS infrastructure to support the product launch. To make the website deploy safely we created Virtual Private Clouds in two different AWS regions with multi-AZ configuration.
  • Setting up Auto-Scaling infrastructure
We created auto-scaling policies based on CPU utilization and inbound traffic using AWS – Elastic Load Balancing which automatically distributes the incoming traffic across multiple targets.
  • Selecting effective Docker Orchestration tool
After exploring many options we selected Amazon ECS due to the flexibility it provides with simple API calls.
  • Deploying Performance Management system
It was important to know if the application was responsive and effective enough, we created a mechanism which would allow us to monitor the system. Alerts were triggered whenever portal became unresponsive or when it produced too many errors.

What we achieved using DevOps

Continuous Integration

Continuous Monitoring

Continuous Deployment

Communication and Collaboration

High Scalability

Robust Infrastructure

UI/UX Upgrades

We incorporated changes to build  superior ‘user experience. We made the application process interesting and exciting for candidates. We introduced Chat based application process where candidate would interact with a Chatbot. The bot, backed by AI and Machine Learning which is capable of formulating questions based on answers given by the candidate in the preceding question like ’,‘Would you be willing to relocate to New York for this opportunity?’ and so on.
We further revamped the UI/UX and made changes in the candidate portal by introducing JD 2.0. JD 2.0 is an innovative way to access the job details page. Employers can easily configure the new branding page or they can even use the current details page. Another striking feature we added is the ability to provide suggestions and recommendations on the dashboard. People searching for jobs now can get suggestions based on the type of work they are interested in and for the employers Shortlist suggests candidates.
In order to incorporate new functionalities swiftly, we used ‘Web Sockets’, which are thin, a lightweight layer above TCP. They make it suitable to use “subprotocols” to embed messages.
JD 2.0 interface
Job Suggestion page

UI/UX Comparison

Version 1

Version 2

Old Sign-up Window
NEW SIGN-UP WINDOW
Old Candidate Dashboard
New Candidate Dashboard
Old Chat Interface
New Chat Interface

The End Result

The upgrade turned out to be a boon for ‘Shortlist’. It didn’t just boosted the performance of the portal but it took the performance of the business to whole other level.
(Visit the new site here : ‘Shortlist)