Jobly - The story of creating our own Chrome extension

Overview

I work in the Upstaff team, where we develop both turnkey projects and help  clients expand the existing teams, i,e, startups, SMBs, delivery and engineering organizations in established businesses. Many businesses want to be able to upscale and downscale promptly, and do not necessarily enjoy waiting 1-2 months while looking the right talent, which is an average time to hire metric across IT industry. Expertise gaps in projects are rather common issue and at some point, needs a solution which is reliable, quick and with an easy engagement. Like, say, call a taxi.

We constantly need help with three main challenges: 

1. Analyze an IT position (especially a large one with several pages), 
2. Get recommendations on job descriptions to attract more professionals as quickly and fully as possible. (We’ve been using Upstaff’s checklists proven over the years, as well as some AI assistance).  
3. Form a short-list of several candidates that fit to the maximum extent, in order to have a choice and options to select from.

Previous history

We already had experience developing Chrome Extensions for our customers in the previous 5 years – unfortunately, I cannot name names due to NDA, but these are mainly investment funds and other companies for whom it is important to see and enrich aggregated information about other companies, their sites, traffic, and internal data. We understood the nuanced complexity and it was a familiar technology for us.

Platform Choice

• Desktop vs Mobile?
  Mobile applications lose to desktop apps, not provide the required speed, and functionality in the recruitment business (talent sourcing, resume profiling, etc. And I am not talking about ATS for now). It’s rather convenient to check something when you travel, chat a few lines with someone, and forward links but not more. Most activity is done in browsers and desktop apps on a notebook. So browser extension is exactly what we needed, and we don’t care if it’s not supported on a mobile version. 
• Chrome, Safari or Firefox?
  It’s a natural choice. We use Chrome as our preferred browser for everyday work, so the extension was developed specifically for desktop Chrome with ability to use on mobile. 

Tech Challenges

• Performance
  The first thing we encountered (including on the example of similar extensions).  Most of the extensions that analyze the content of the page and make changes (for example, highlight elements) begin to make themselves known and slow down the browser by increasing the number of tabs where the extension works. For example, when you have several or more bookmarks with target resources such as LinkedIn – that’s all, work becomes impossible. When we install several such plugins – the situation becomes even more complicated. We had cases when we simply could not do anything because the tabs froze and stopped responding. And this was from the known And it is even more difficult to understand – which of all the extensions is stuttering more than the others and deactivating it.
  Therefore, the first thing is performance. We open what we studied at the University, the complexity of algorithms, big O, etc. and find a solution. The document is parsed at an acceptable time, and technical skills are highlighted. The result satisfies us.
• Display Options
  We played with drop-down windows but stopped at the sidebar because the windows constantly close when you go to another page or just a careless click. Also, it is unclear whether there are infies or not if the extension is closed. Therefore, our choice fell on the sidebar, which Google developers provide in the tools. Not immediately, but it worked as it should.
• Web resource coverage & support
  We proceeded from the reasoning that 90% of recruiters use similar platforms for posting jobs (Indeed, Glassdoor, Linkedin Jobs, Workable, JobLeads), and we implemented support for these platforms first. Abstract parser of abstract pages, where content may occur – the plugin may not correctly highlight technical terms, or find elements. We are considering supporting Google Docs (as an interim option for preparing a job description), but for now, we have stopped at the platforms themselves. Of course, if the page templates of these platforms change, we may need to fix something ourselves, but this does not scare us. We desire to support the plugin in the future.
• Text analysis within job and resume context and evaluating both for the best match
  The context of parts of words in a sentence, for example, the programming language Go / Golang. People often use the verb go not at all as the name of a programming language, or the official library next (next.js). This also raises some questions, but the issue is solvable. Where should the analysis be performed, and with what tools? As a front-end, we chose javascript and vue.js (3). This allows you to work with the DOM of the document and make the necessary changes to the page. Regarding analysis, there are not enough libraries and frameworks for working with text arrays on the front end, it is more about backend tools such as ElasticSearch or AI models. Therefore, we decided to bring the analysis that is possible to do to the front-end and move the rest of the analysis to the back end, and make it available for public use in the Pro version.

Features Overview

I suggest going directly to the functionality overview. It was designed to be simple for the end user.

1. As soon as you enter one of the popular platforms such as Linkedin, Indeed, Glassdoor, or Upstaff, you get an X-ray in terms of technical requirements: programming languages, frameworks, libraries, utilities, small tools – all this is highlighted on your page.
2. In the sidebar, you will receive this information in a structured form, grouped by categories. For each name, you have indicators of popularity in the industry, and how much this term is disclosed in the job description. Thus, we have indirect information that the author of the Job Description considers more important, and how much this correlates with the general “temperature in the hospital” / industry. You also get an assessment of whether enough technical terms are used in the job description.
3. In the next block, you will see suggestions for using technical terms. For an IT specialist, it is important to get a description of their own system of coordinates and terminology. We also take into account frequent misspellings, abbreviations, and long names and offer alternative options if they are more familiar to developers.
4. At the top, you see the overall Job Description Score. The goal is to achieve the maximum score. This means that within the scope of the job description, the work is done, and now you can proceed to posting on other resources and ATS.

The outcome: “Jobly by Upstaff” Chrome Extension

The extension we created helps users identifying best matching candidates for job postings.

Download Jobly by Upstaff, free online job scanning and candidate matching tool from the Google Chrome Store.  

If you read this to the end and find it relevant, I wish you success, and see the best developers in your dream team.
And you’re welcome to share your project or product requirements with Upstaff and get a free quote, and engage expert-level engineers into your project within days. As a partner in AI, Data, Web, and Mobile, we will help you implement the project or hire developers, who are able to solve any challenge.