Purpose-Built Machine Learning: Smart Shortlisting Fueled by Diverse Data

Share This Post:

Purpose-Built Machine Learning: Smart Shortlisting Fueled by Diverse Data

Reading the title of this piece, you might be wondering what we’re talking about – we’ve packed a lot of big concepts into only nine words. But, at a high level, what we’re talking about is hiring smarter, about applying intelligence to the recruiting process to help you discover relevant candidates while reducing the risk of bias. These are important ideas, ideas that are shaping the future of hiring. So, if you’re in the business of recruiting, or interested to understand what we’re building at myInterview, keep reading. 

Let’s start by breaking down the title piece by piece. Then we’ll put it all back together and look at what it means for the day-to-day of recruiting. 

Purpose-built 

Purpose-built refers to a specific kind of solution. In this case, we’re using it to modify machine learning. You might also see purpose-built ahead of artificial intelligence or other types of technology. As the name indicates, purpose-built means the solution was designed to support specific requirements instead of an all-or general-purpose solution. The difference being, all-purpose offers a one-size-fits-most approach that requires heavy customization, whereas purpose-built provides easy integration with other systems and the flexibility to scale to your business needs.  

Purpose-built = solution with an explicit reason for being

Machine learning

Machine learning (ML) is a broad term that typically sits under the artificial intelligence umbrella. But machine learning refers explicitly to computer algorithms. These are models built on sample data that improve through experience and use, learning along the way. In talent acquisition, ML algorithms can be used to interpret body language and automated transcriptions. Doing this can better an interviewer’s understanding of a candidate and improve hiring outcomes. There’s also the famous quote from MIT research scientist Andrew McAfee, “If you want the biases out, get the algorithms in,” because the knowledge we uncover through ML can identify patterns, reduce biases and assist in decision making. 

Pausing here for a moment, if we put purpose-built and machine learning together, we get algorithms designed for a particular reason. In our case, that’s interviewing, which leads us to the next part. 

Purpose-built + machine learning = algorithm-based solution with an explicit reason for being 

Smart Shortlisting 

Smart shortlisting is a feature of myInterview. It’s a functionality we created to help recruiters get additional insight from video interviews and determine which candidates are a match for the job opening. Here, the algorithms review each video and analyze candidate responses for soft skills, personality traits, and keywords. In turn, recruiters are able to see automatically which candidates are most relevant and prioritize them. So, it’s not a question of ranking candidates so much as it is re-ordering.

By now, you can see we’re beginning to put action behind intent, but there’s one more piece to consider. 

Purpose-built + machine learning + smart shortlisting = algorithm-based solution with an explicit reason for being able to highlight perfect fit candidates 

Fueled by diverse data 

We established above that ML algorithms involve data. What we didn’t explain was where that data comes from and how it gets used over time to refine the algorithm and enhance the solution’s understanding of candidates. When we say fueled by diverse data, we mean that myInterview baked diversity into the development of the algorithm to mitigate the risk of bias. The more diverse your data is from the start, the more balanced the algorithm. But it’s important to remember that the algorithm isn’t a determinate. It’s a tool developed for the express purpose (there’s that word again) of facilitating the hiring process. We are fueling this process with diverse data, making sure everyone benefits.

Purpose-built + machine learning + smart shortlisting = algorithm-based solution with an explicit reason for being able to highlight perfect fit candidates while mitigating bias through varied datasets 

The benefits 

We want everyone to have a fair shot at getting hired, and so we created myInterview to make that vision a reality. And going back to our title, you can see how our concept becomes easier to comprehend. We’re talking about intelligent technology, built with a specific use in mind, that includes a diverse data set and makes it possible to recognize which candidates to advance first. It’s machine learning that enables people to do their job better, not do their job for them. At the same time, this approach gives candidates the ability to showcase their skills, talents, and abilities, free from bias. 

Seemingly abstract ideas like purpose-built, automation, and machine learning have a tendency to scare people at first, unsure of the meaning and application. But, when we dig in, it becomes clear that these words and phrases are the distillation of practical uses that ultimately make our life and work easier. When it comes to recruiting, there are a lot of buzzwords that get bandied about, taking away from the central point. We want to focus on what really matters, and that’s finding those hidden gems, the candidates who turn into great hires and help drive your business forward. 

About the Author

Clayton Donnelly

Clayton Donnelly is Chief Behavioral Psychologist at myInterview.


Did you find this interesting? Share it with others:

POPULAR POSTS