HackerEarth

author’s Articles

Insights & Stories by Keerthi Kumar

Whether you're building your first team or scaling culture across regions, Keerthi Kumar's articles offer human-first insights rooted in real practice.

AI Recruiting

6 Ways Candidates Try To Outsmart A Remote Proctored Assessment

Tech hiring has undergone a radical transformation with take-home assessments gaining popularity over traditional methods in recent times. Since even a single instance of candidate malpractice can taint the efficacy of the entire test process, it stands to reason that take-home assessments and remote interviews would never have gained their current credibility without a foolproof online proctoring mechanism.

Proctoring is crucial to validate a candidate’s authenticity and integrity. Without some sort of online supervision, it would be impossible to claim that assessment platforms such as HackerEarth offer accurate results and candidate analysis.

In short,

Standardized Developer Hiring = Customized and Intuitive Coding Tests + Robust Proctoring

So, without further ado, let’s take a look at the 6 most commonly deployed tricks your assessment tool needs to guard against, shall we?

#1. Switching tabs to get online assistance/reference during a test

El switch-a-tab-aroo.

Switching tabs to get online assistance/reference during a coding test. Proctoring tips on how to deal with such a case

This is a classic one, and probably the most obvious example in our list. A lot of candidates happily assume that interviewers and creators of assessment software blindly trust in a candidate’s ready inclination to always take the moral high ground. But the angelic souls who create such software are well aware of this age-old technique and have hence installed divine software intervention to guard against it. #NotAllDevelopers

Fixes: Full-screen mode, Custom timer on MCQs, Automatic log-out.

#2. Copy-pasting code from another computer or from the web

El copy-paste-o.

Copy-pasting code from another computer or from the web. Proctoring tip 2 on how to handle this

Another age-old trick in the digital book is the incredibly humble copy-paste. With the latest upgrades in Windows and Mac operating systems, candidates can, in fact, store and keep-at-the-ready, not one but a string of copied texts. You can access these copied pieces of code with just a couple of keystrokes. This includes productivity tools, messengers, data and word processing apps, Mac cleaners, and security software.

Fixes: Copy-paste lock, Full-screen mode.

#3. Getting other users to take the test on their computer

El Imperson-anaa-tor.

How proctoring can detect other users trying to take the coding test on their computer.

What’s the best part about take-home coding tests? Well, it doesn’t necessarily have to be you that’s taking them. Your coding prodigy of a best friend, or perhaps that bespectacled kid in class who hibernates in the library, could take them for you, all for a quick buck.

Fixes: Randomized webcam snapshots, IP Address Lock, Automatic impersonation detection.

Also read: Importance Of Online Proctoring In University Hiring

#4. Moving away from the webcam to look into another laptop/screen/book.

El dos screen-o.

Moving away from the webcam to look into another laptop/screen/book during a coding assessment.

A lot of candidates, unfortunately, give in to the overpowering urge to sneak a peek into another screen or a book. They’re often prepared with a secondary device that’s kept hidden from the web camera’s line of sight. So, it’s extremely easy for them to access these hidden resources unbeknownst to the system. What’s more, the screens of neighboring candidates are also a potential cause for concern.

Fixes: Randomized webcam snapshots, Plagiarism checker, Automatic mobile phone detection.

#5. Having another person in the room who can read out answers and solutions for you

El dos candidate-o.

Having another person in the room who can read out answers and solutions for you (during a coding test).

A good friend will bail you out of jail. But a best friend will be sitting in the cell with you, going “Wow, wasn’t that something?”.

God bless friends who will, more than willingly, aid and abet candidates in a malpractice attempt. Candidates take advantage of the fact that there aren’t any audio recording devices embedded in the test system. So, they can either try taking the assistance of a gifted comrade in the case of remote assessments, or others who are also taking the test if using an onsite group interview room.

Fixes: Randomized webcam snapshots, Plagiarism checker.

#6. Restroom breaks

El missing suspect-o.

The final and the most elementary (quite literally) of all is the innocent nature’s call card. In the case of remote examinations, the computer and its camera basically account for all of your exam halls, and everything that’s out of the webcam’s line of sight is not. For the briefest time, you’re Alice in ‘Wander’land, and the whole house is your deception oyster.

Fixes: Randomized webcam snapshots, Custom timer on MCQs.

Also read: 5 Must-Have Proctoring Tips For A Developer Assessment Platform

9 Ways They Aren’t. (HackerEarth Proctoring Fixes In Detail)

If you were worried about what you read above, then allow yourself to breathe a sigh of relief. HackerEarth Assessments has in-built proctoring features to ensure that every assessment taken on our platform is monitored to ensure authenticity. Here are the most important ones:

#1. Randomized webcam snapshots – 2 every minute

This is the biggest and most potent gun we have in our arsenal. The reason this ranks the highest in our list is that, unlike the other features which are announced by the system, it is easy to forget about this functionality. Randomized webcam snapshots basically pick up any sort of malpractice attempt, ranging from peering into a second screen to verbal communication with another person in the room.

#2. Custom timers on multiple choice questions

This is based on the premise that it takes time to look for or seek any form of assistance. So, a timer on the MCQ questions is a good way to keep the pressure on the candidate, discouraging them from wasting any time scouring the internet or elsewhere for a solution. The other logic behind this feature is that the short amount of time will simply not allow you to cheat, as it just won’t be humanly possible to look for hints, do your calculations, and finally punch them in.

#3. Plagiarism checker. All submissions need to be unique

Another heavy piece of artillery in our war against interview malpractices is our system’s inbuilt plagiarism checker. This advanced AI-assisted function is every recruiter’s dream come true. It investigates a candidate’s code against all the other applicant submissions for the same test.

Even the remotest match in code is picked up by the system immediately and the respective administrator is promptly notified.

Also read: 3 Things To Know About Remote Proctoring

HackerEarth add-ons for robust proctoring

Here are some additional proctoring fixes that further bulletproof our assessments, making it one of the most reliable and accurate sources of remote technical candidate evaluations out there in the market today. Learn more about them in detail in our support article, here.

#4. Restrict test access for certain IP addresses.

#5. Disable the Copy and Paste feature in the code editor.

#6. Negative marking could be activated during MCQ questions.

#7. Automatic impersonation detection (plus) mobile phone detection.

#8. Automatically log a candidate out upon leaving the test environment.

#9. Restrict candidates to full-screen mode for the entirety of the assessment.

Also read: Screen Candidates Better With HackerEarth's Proctoring Feature

Automated invigilation made easy with HackerEarth!

Like with all regulatory orders found in society, there will always be attempts at finding new and innovative ways to beat the system. That’s just natural human behavior, and we fully condone it.

The way we look at it - it’s really a game of who’s the smarter fox in the jungle. In order to help recruiters stay on top, we constantly update our system to stay ahead of the curve. Our team vigilantly looks for dodgy trends much ahead of the competition and anticipates candidate behavior to create smarter features.

HackerEarth has a 7-million-strong developer community, and we know how easy it is for a few bad apples to give the entire community a bad rap. Our goal has always been to create a system that provides developers with fair and objective assessments and helps recruiters pick the best talent. The proctoring features in our Assessment platform do both with equal ease.

Subscribe to the HackerEarth blog and enrich your monthly reading with our free e-newsletter - Fresh, insightful, and awesome articles straight into your inbox from around the tech recruiting world!

Developer Insights

HackerEarth Introduces Full-Stack Assessments At Hire10(1) Conference

The new HackerEarth Assessments feature of full-stack assessments will facilitate both backend and frontend developer assessment for recruiters.

HackerEarth has just announced the addition of full-stack assessments to help recruiters efficiently evaluate the coding skills of full-stack developers. HackerEarth’s CEO, Sachin Gupta, made the announcement today at Hire10(1), HackerEarth’s flagship virtual conference to help recruiters and engineering leaders hire top developers and build great tech teams.

According to the 2020 HackerEarth Developer Survey, more developers — over 35% — have expertise in full-stack development than in any other category. Yet evaluating these skills is notoriously difficult because full-stack development spans across multiple skills and requires a high level of customization based on the specific technology stack that an organization uses. HackerEarth’s full-stack assessment solution is highly flexible and customizable, supports a large number of out-of-the-box tech stacks, while simultaneously allowing any custom stack to be installed in the development environment. These assessments also include a powerful browser-based IDE built on top of Theia editor, providing developers the same code writing experience in the browser as they would get on their own systems.

“Organizations are increasingly looking to recruit full-stack developers. In fact, recent data from a survey done by Indeed shows that the demand for full-stack developers in the U.S. increased by 206% between 2015-2018. However, finding a good full-stack developer is hard since the assessment process is notoriously difficult and can take hours, days, and even weeks” said HackerEarth CEO, Sachin Gupta. “We added this feature to simplify the process and are working to make it easy for recruiters to proctor over longer time periods and assess using various programming languages. Our goal is to enable recruiters to evaluate developer skills from early candidate screening to more in-depth assignments in the later stages of recruitment, as well as to provide the best performance and experience possible for full-stack candidates.”

Key Benefits of the HackerEarth Full-stack Assessments Include:

Flexibility: Can be used for full-stack, or for assessment of either frontend or backend skills independently
Breadth of Languages Supported: Includes Python, Django, Java (Spring), Node JS for backend, and React and Angular JS for frontend
Customizability: Recruiters can customize the development environment and the task based on their specific technology stack. Candidates can then build and run entire applications within the HackerEarth portal
Ease of Implementation: Gives full access to the HackerEarth library of 13,000+ questions which allows recruiters to quickly build out full-stack assessments or create custom questions as well
Automation: Fully automated backend assessment, with frontend coming soon

For a better review process, the product has a detailed report section with additional functions including:

Real-time recording of actions taken while building the application in the form of log files which recruiters can download and review
A preview function to help recruiters and candidates check the build easily
Automated screenshots of the application built by the candidate so recruiters can quickly evaluate their progress

Features Coming Soon:

Automated code quality score
Fool-proof proctoring for longer-term assignments
Support for more languages and frameworks, including PHP and Ruby on Rails
Auto-evaluation for frontend
The full-stack infrastructure will also be used to create specialized assessments for roles like Cybersecurity engineer, Game developer, etc.

Learn more about HackerEarth’s Full-stack Assessments.

DEI Hiring

This is Recruiting - Demystifying bias in recruiting and how to tackle it.

Welcome to another interesting episode of "This is Recruiting", a series that equips HR professionals and tech recruiters across the globe to gain actionable insights from fellow recruiters to take their hiring to the next level.

In this episode, we caught up with somebody special, someone with a gold mine of useful information regarding technical recruitment. David Windley, CEO, IQTalent Partners, who is also Board Chair for the Society for Human Resource Management (SHRM) shares with us a generation's worth of recruiting wisdom and valuable insights that he's picked up over the decades. Having spent around 30 years in corporate HR, David is one of the leading industry experts in the world of recruitment. From all his years of observing, dealing with, and building processes around bias in hiring, he has much to say and offers us timeless advice on some of the best ways to tackle it.

The first step is always to call it out, he says. It begins with acknowledging that bias exists, then by rooting out the bad biases that aren't performance-driven out of the process, and lastly by building workable systems around that.

He maintains that the only way to overcome bias is by having recruiters zero back to the original principle of assessing the individual based on merits alone, remembering that they need to have the best interests of the broader organization in mind, and not give in to their personal inhibitions and prejudices.

This is Recruiting - Reducing bias in the hiring cycle.

Sachin:

In your opinion, how important is it for an organization to focus on reducing bias while hiring?

David:

So, let's set aside the social issues. There are reasons to do it because of the broader social good. But let's just talk as a business.

Our goal when we're trying to hire people is to really find the right people that will be the best performers in our organisation - as an individual and collectively within our culture and company. So, when we're trying to find the right characteristics that will lead to good performance and when we have bias creep in here - it's only going to hinder our process of finding the ideal candidate for the position.

Bias that's unrelated directly to performance will only cause you to sub-optimize in your decisions. From a pure business perspective - all of us should want to address this issue.

Sachin:

I'm sure you would have seen the length and breadth of different organizations and the functions within. In your experience, do you see certain functions that tend to be more diverse? Or the converse of that?

David:

Yeah. Depending on how you talk about diversity. There is ethnic diversity, there is gender diversity, and then the broadest of all - diversity in thought and perspectives. But, yeah. If you just look at demographics and statistics, there are certain functions that lean more towards certain gender demographics and also ethnic demographics. So that's true.

Again, that doesn’t necessarily mean that we should just then assume -- because at a macro level -- those statistics are what they are. That means anything for any individual.

So going back to the first question. A very good example of how bias creeps in is when someone looks at a macro and just makes an assumption based on ‘association by group’. But how much of those macro statistics have bias built into it is due to maybe reasons like bias in society, etc. So yes, at a macro level there are just historical differences in certain functions. The point is for any individual that you are assessing, you are trying to discern that person's capabilities, skill sets, and competencies; whether they're going to be a good performer and fit for your organization.

The Go-Getter’s guide to diversity hiring in tech

Sachin:

Considering that humans are hardwired to align with people similar to themselves; affinity bias is so hardwired into us that it isn't that easy to overcome. So, in such a situation what are your guiding principles that help you make the right decisions in the recruiting process? And what have you done with your team over the years?

David:

Yes, I think you make a really good point. That's where we start with this issue on bias - to understand that it is natural for humans to categorize. That's just how our brains work.

There is just so much information out there that we have to categorize things and it's how we work. We need to just realize that bias is a natural thing and that we all have biases.

We all hear messages, we grow up in our societies, and whatever messages or things we learn or observe in those societies, they enter our unconscious and conscious mind.

So, let's first just demystify it. Bias exists. And the first thing to do about it is admitting that that's the case. The issue now is to deal with the unrelated biases and to get that out of the process, so it doesn't get in the way.

Why do I say that? Since there are obviously some good biases too. For example, I have a bias for people that are self-starters. I think that's an okay bias because it's performance-related. But having a bias about someone's gender, or someone's ethnicity, or race is not directly related to those sorts of performance behaviors. So, from a process point of view,

It's good to have a structured interview assessment process that identifies the characteristics and competencies that you're looking for.
Having structured questions around that and having a nice feedback loop as a team to make sure that when you're assessing, you are, in fact, talking about those characteristics.
Not relying on the shorthand - "Joe is a good guy. I like Joe." That is not a good assessment. That doesn't work.

Want to keep going? Sachin and David go on to talk about centralized recruiting teams, the role of AI in reducing bias, hiring patterns and outlier statistics, diversity training, and more.

Listen to our entire conversation with David here.