A tech recruiting platform used by Amazon and PayPal is tackling bias by eliminating the need for resumes and cover letters in the hiring process — here's how it works
- While women make up nearly half of the overall US workforce, they hold less than half of the positions in the top US tech companies. The numbers are even lower for people of color.
- Recruiting firm HackerEarth is hoping to solve the diversity problem in tech with its new anti-bias platform.
- They've restructured the hiring process so that candidates are evaluated solely based on their skill level through at-home, blind assessments — all other details around their gender, race, or educational background are removed from the recruiter's view.
- "We have seen a lot of interest and adoption from our enterprise customers," said Sachin Gupta, founder and CEO of HackerEarth. "Fighting bias is a much more acute problem for larger organizations because it can be difficult to monitor unconscious bias across a larger panel of interviewees, hence there is a need for a solution that is embedded into their tools."
- Visit Business Insider's homepage for more stories.
While many high-paying fields like medicine and law pose plenty of barriers to entry, including the number of years of study and degrees required to practice, another lucrative industry — the tech sector — offers the allure of an easier path. At least, on the surface.
"Today, anyone with a laptop and internet connection can learn to code," Sachin Gupta, founder and CEO of HackerEarth, a global company that connects talent with opportunities at Amazon, PayPal, Intuit, and more, told Business Insider. "It doesn't matter what your educational background is or where you worked in the past. The only thing that really matters is your skill."
The problem is, while there may be fewer challenges when it comes to acquiring the skills needed, the tech industry is not without its own barriers to entry. After all, the field has a well-known diversity problem when it comes to hiring equitably across gender, race, and ethnicity.
While women make up nearly half (about 47%) of the overall US workforce, they hold less than half of the positions in the top US tech companies, making up 29 to 42% of the general workforce at Microsoft, Google, Apple, Facebook, and Amazon — and face even worse odds when it comes to leadership roles, holding 20 to 27% of these positions among these same employers.
Women of color face even greater challenges, with about 13% of computing and math-related jobs held by Black women, 6% held by Asian-American women, and 5.4% held by Latinx women — compared to 25% of computing-related jobs held by women of all racial and ethnic backgrounds, according to the National Center for Women and Information Technology 2020 Scorecard.
In the fall of 2019, Wired reviewed the first five years of tech diversity reports from Apple, Facebook, Google, and Microsoft and found that while the share of women in tech has made some limited gains, there has been little growth for Black and Latinx workers in tech — and the industry overall is still disproportionately comprised of white or Asian-American male workers.
"The problem of unconscious bias is real and deep rooted and we need a systematic approach to solve it," said Gupta.
That's why Gupta founded HackerEarth, a tech recruiting firm that's helped place hires at over 1,000 companies including Walmart Labs, Wells Fargo, and Barclays with the goal of making opportunities in the tech industry accessible to candidates of all backgrounds.
And there's a real need for the work Gupta's team is doing, as unconscious bias in the hiring process is a well-studied problem, with candidates' gender, race, and ethnicity known to influence the decision-making process of recruiters and hiring managers.
Since its creation, HackerEarth — which offers companies tools to offer candidates at-home and live coding assessments as well as conduct online video interviews — has focused on making the hiring process more objective by helping recruiters and hiring managers select candidates based on skills more so than any other criteria.
To take things a step further, HackerEarth recently launched a pilot of a new candidate screening process to help curb the problem of bias right at the very start of the recruitment funnel.
"Applicants who progress to a later round of interviews [are] generally subject to less bias because by then they have already demonstrated their capability," said Mitali Sodhi, the lead data scientist responsible for launching HackerEarth's anti-bias feature. "So by putting the anti-bias process right at the front of the recruitment funnel, we can make a significant impact on the overall process."
How it works
To date, HackerEarth has screened over four million developers. When connecting with current and potential customers — many of whom were implementing diversity and inclusion initiatives to help fight bias within their workforces — the team saw an unmet need: To completely overhaul the candidate screening process to remove personally identifiable information (PII), such as individual names, birth dates, genders, and racial and ethnic identities, from the assessment process.
"After speaking with some of our prospects who were not using any remote assessment tools, we realized they wanted to move away from using resumes as a signal for skill but they did not have alternative tools to measure competencies," said Sodhi. "Based on this feedback, we realized that if we were able to mask PII during our remote assessments without compromising the quality of skill evaluation for candidates, we would have a potent solution to the problem."
With its new anti-bias screening technology, launched earlier this year, employers can use HackerEarth's platform to screen candidates in two key ways that both work to eliminate the need for cover letters and resumes while also reducing the impact of hiring managers' personal prejudices on the selection process.
With the first step, every single potential applicant for a given job opening — except those that may be disqualified based on factors like their location or not meeting work permit requirements — automatically goes through a pre-screening process with the chance to complete remote, at-home assessments.
"These take-home assessments are hands-on programming tests that are specific to the skills that the prospective employee needs to have in order to perform their role on the job," said Gupta. "For instance, if you are hiring a backend developer who has to write code in Python, the assessment will have programming tasks in Python that simulate their actual on-the-job work."
Next, hiring managers review the performance of each job applicant and are then prompted to select a shortlist of candidates to advance on to the next step in the process — without seeing any PII, such as a candidate's address, email address, educational background, or employment history. Instead, all they see is a randomized and anonymous ID for each individual. About five to 30% of candidates usually advance from this first stage to the next round, Gupta explained.
Candidates that make the shortlist are then invited to take part in a pair programming interview, where a member of the hiring team and the candidate collaborate on coding tasks together remotely.
"Even for the interview, the interviewing panel is not shown any PII and the video cannot be turned on during the interview," Gupta said. "It's an audio conversation where the candidate is given actual coding problems to be solved in real time."
There may be as many as two to three rounds of these remote interview assessments, with the first round screening out another 20 to 40% of candidates, Gupta explained. By the final round of assessments, hiring managers usually have enough information to make the final offer, though some customers conduct final HR interview rounds outside of the HackerEarth platform.
While resumes and cover letters are not typically a factor during the initial candidate pre-screening process, Gupta said some companies may choose to prequalify candidates with a few questions, such as if they're willing to relocate, if they're permitted to work in the US, why they're applying for that specific job, or what recent projects they've worked on.
Sakshi Jain, a data analyst who was hired at HackerEarth earlier this year through this pilot screening process, attributed the system to her being considered for the role, despite having come from a different background as a systems engineer and having a three-year gap on her resume.
"The fact that I didn't have prior experience in the specific role I was applying for and was in between jobs didn't put me at a disadvantage," she said. "The decision to go forward with interviewing was based on my remote assessment and interviews that primarily focused on the skills needed for the job."
Rolling the platform out and planning future iterations
At this stage, HackerEarth is collecting feedback about the technology directly from clients and tracking metrics related to feature adoption and usage to gather insights about how it's been impacting the hiring process. Once they have enough data, the plan is to roll this feature out to their wider customer base.
"We have seen a lot of interest and adoption from our enterprise customers," said Gupta. "Fighting bias is a much more acute problem for larger organizations because it can be difficult to monitor unconscious bias across a larger panel of interviewees, hence there is a need for a solution that is embedded into their tools."
Since it's only been a few months since the company initially launched the pilot, HackerEarth is not yet able to calculate the effect on any diversity and inclusion hiring outcomes for its clients.
Before piloting the technology with customers, however, HackerEarth conducted an experiment on a group of 300 people from different socioeconomic backgrounds and evenly split between males and females to see whether there were unintended (and potentially harmful) consequences of using the assessment tool as a screening process.
"The tests conducted on that control group demonstrated that there is no adverse impact of using our assessment tool," Sodhi said. "In fact, the tool is perceived to be positively biased towards lesser represented groups because it levels the playing field for everyone."
In addition to removing requirements like education from certain types of schools or experiences at certain types of organizations from its own job postings, the company has adopted the technology for its own hiring efforts.
Another feature the company is exploring includes voice modulation to help minimize the effects of candidates' speech on the decision-making process.
One limitation the company recognizes is that technology can only do so much.
"The fight against bias is a hard one," said Sodhi. "This also requires a strong commitment within an organization to train its employees on what unconscious bias is and ways to reduce its effects during the hiring process."
Source: Read Full Article