Climb Credit, an education lender focused on private vocational schools, is partnering with software provider Zest AI to bolster its credit offering with credit underwriting powered by machine learning (ML); the partnership was announced last week.

The lender will use Zest’s model management system to power its student lending credit model to enhance accuracy and fairness in its loans. Climb plans to use the software to process credit bureau data and reweight factors that can serve as proxies for race, gender and other protected class statuses, it said in a statement. Climb partners with vocational and skill-based training schools to make loans to students who qualify for courses that have a demonstrated ability to boost student income.

It was “frustrating to be running a company that relied on FICO [traditional credit scores],” Angela Ceresnie, chief executive at Climb Credit, told Bank Automation News. She added that credit score-based systems tend to, “regress to the mean,” and bake in certain structural biases that can deprive people of access to credit.

Under the partnership with Zest AI, the lender will use machine learning (ML) techniques to underwrite loans and reduce the disparity in credit by leveraging the “adversarial debiasing” method used by Zest.

“The current system is broken,” Mike de Vere, chief executive at Zest AI, told BAN, adding that lenders have traditionally used the “drop one” model to measure and control for bias in metrics. Under this approach, lenders add or remove factors to see whether they indicate protected class variations in credit.

Adversarial debiasing, on the other hand, uses “two [machine learning] models that learn together,” de Vere said. The assessment model, calculates the risk associated with a loan while the coaching model is constantly “nudging” the assessment model to check for bias and be fairer, he added.

Other fintechs have also recognized the student lending space as a viable growth area, with CommonBond, SoFi, Credible and Earnest all expanding to help students refinance loans at lower rates or take out their original loanes at lower rates.

“Climb’s loans are in -part backed by Goldman Sachs,” Ceresnie told BAN. While she declined to specify the volume of loans made by Climb, she said she expects larger institutions to also leverage advanced techniques like artificial intelligence and ML going forward, although they’re currently held back by their size and outdated tech stacks. She said partnerships with software vendors are likely to allow big banks to expand into student loans.

Tech Crunch previously reported that Climb Credit has contracted with an “undisclosed top-three global asset manager,” to secure a lending facility of up to $400 million to make its loans. The lender has raised a total of $68.6 million in funding, and Zest AI has raised $80 million in funding from investors like Insight Partners and Chinese tech giant Baidu, according to Crunchbase.