WorkClass

Full Stack Developer

WorkClass is a job portal that is aimed at helping blue collar workers find jobs within 24 hours. Over 5000 employers use WorkClass and it boasts more than 60,000 monthly visitors on its jobseeker website. At WorkClass, I worked as a full stack developer, maintaining the various products and websites used by WorkClass' jobseekers and recruiters.

 

WorkClass had placed a huge emphasis on improving the website's organic traffic using Search Engine Optimisation (SEO), and I played a large role in helping to achieve this goal. This involved changing and improving on multiple aspects WorkClass' frontend, such as configuring static-site generation (SSG) or server-side rendering (SSR) for our most visited web pages and identifying and fixing issues affecting Google's Core Web Vitals (CWVs).

 

In order for WorkClass to stay relevant for certain keywords on search engines, I had also generated backend scripts that automatically generated blogs for our most popular keywords. This process involved gathering the our most popular keywords from Ahrefs and then using OpenAI's GPT-3 language model to generate possible blog titles. From these blog titles, I used GPT-3 again to generate blog section subheadings, and then finally written content for each section. This script has helped our in-house copy writers push out content faster than ever before, as they need only to vet the blog content before pushing it to the frontend as a new blog.

 

During my 9 months at WorkClass, the SEO improvements that I made benefitted WorkClass tremendously and the overall organic traffic to its Jobseeker website had more than doubled.

 

Apart from SEO, WorkClass was also heavily focused on improving product usability and features for recruiters. To that end, I was also heavily involved in implementing new features for WorkClass as well as maintaining old ones and updating the UI. Some features that I had helped to implement/maintain were search functionality for jobseekers and recruiters, interview scheduling as well as autogenerated screening questions based on job descriptions set by recruiters. The list mentioned above is non-exhaustive and too geeky to explain in-depth in a single write up, however if you are still interested to know more, do email me at aikenwongx@u.nus.edu.

 

At WorkClass, I had also helped to set up and improve numerous continuous integration and continuous delivery (CI/CD) workflows. This involved setting up parallelized cypress integration testing for our frontends using Github Actions, automatic reporting and screenshots during test failures as well as scheduled cypress testing on our production sites within a docker container. These workflows had helped to speed up the pace of development for the entire team and afforded us the assurance that any successfully deployed code worked correctly.