The modern job hunt requires deciphering job descriptions that often read like secret codes. Job seekers are tasked with interpreting these codes and translating their own experiences to fit the qualifications of new, rapidly changing positions. UpScored is working to ease this process with a career discovery platform that leverages a natural language processing and machine learning search algorithm to connect job seekers with opportunities suited specifically to their skills.
The UpScored job-matching algorithm accounts for candidates’ individual skills, work experience, and education to develop personalized job recommendations. Users receive search results with a compatibility rating on a scale from 1-100, which reflects how close of a match the candidate is for the job. The algorithm’s machine learning approach enables adaptive search returns, so each new batch of matches reflects a user’s preferences.
Co-Founder Elise Runde Voss explains the motivation to build UpScored came from a particularly challenging experience with recruitment after Voss and her future co-founders were charged with launching a data strategy group at SAC Capital. Tasked with hiring nearly 25 people for data science and business development roles, they quickly found that it was hard to predict whether a potential candidate was qualified for a position.
Voss explains, “We were spending hours and hours searching through resumes and going through LinkedIn profiles, it was such an inefficient, messy process, so we basically thought, is there a way we can use data science techniques that we’re using in our current job for investment reasons and apply it to the recruiting space?”
In February 2015, Voss and her co-founders launched UpScored.
“The idea was to take natural language processing and machine learning and look at the relevance of a candidate’s resume versus the job description. It’s almost a compatibility score with work experience, background, skill set and education.”
The algorithm has been trained on almost 500,000 resumes and can match with over 16,000 job descriptions on the site. Thanks to a relevant feedback function, after one round of vetting 25 jobs, users will start to see personalized results. Voss says the function is particularly useful for someone looking to switch careers, since the algorithm can identify parallels in skills across multiple industries.
In the early stages of developing the platform, the team consulted third party recruiters to learn about their experiences with recruitment. From one pilot program, they found that recruiters themselves weren’t always certain about what characteristics the best candidate for a job might possess. Often, it wasn’t until applications started rolling in that recruiters really got a sense of what they wanted. To account for this process, UpScored built responsive recruitment technology.
“Once the candidates start coming through, the hiring manager or internal recruiter can star candidates they like and it will start to look at commonalities across those candidates and it will then surface more people like that to the top.”
Next year, Voss anticipates the release of a career progression function, so a job seeker can see the skill gap separating their own qualifications from what a job requires and what actions could be taken to narrow that gap. Ultimately, Voss says UpScored is on a mission to improve livelihoods.
“I feel like recruiting and hiring are boring words. It’s people-matching. It’s finding a better way to make someone’s life better and help them find their dream job.”
From the markets of Budapest to the skyscrapers of London, Caitlin is our expert on all things European. Located in Copenhagen, Denmark, Caitlin keeps us up-to-date on the European startup landscape, plus culture, food and whatever she finds along the way.