Blinder Resume

Applicant dashboard
Recruiter dashboard
Resume review

Inspiration

Initial resume screening is an important step where recruiters' bias plays a great role in discrimination. Applicants are highly discriminated against based on their race and gender in this step. Applicants with 'white names' and 'male names' are multiple times more likely to receive an invitation for an interview that their counterparts with 'non-white names' and 'female names'. To combat this we were creating a program to anonymize and generalize applicants' resume elements to create a more blind and equal-opportunity resume screening process.

What it does

Based on our research the main identifying elements of a resume for race and gender are name, contact information, and extracurriculars. As such, our program reads a resume and hides name, and contact information from the recruiter. In addition, still in development, it generalizes extracurricular activities in the resume so that recruiters have good information about the applicant's interests and efforts without consciously or unconsciously discriminating against them based on the types and details of the activities.

How we built it

We used python libraries to read and parse the resumes of applicants. Following that we used Open AI's GPT-3 to run NLP and extract relevant information regardless of the formatting of the resume. Following that, still in development, we further utilized GPT-3 to generalize and categorize extracurriculars and other points based on which potential discrimination can occur. We used Django to combine these elements together in a web app. We developed a front end both for the recruiter and the applicant. For the applicant, after submitting information and resume, identifying information is removed from the resume and saved in a separate database where it can be fetched by the program. In case the recruiter wants to select the applicant for an interview, then a simple click can send an email to the applicant.

Challenges we ran into

The major challenges we ran into are time constraints and standardizing the format of resumes, of which there are many.

Accomplishments that we're proud of

Even though we have yet to complete some features, we were able to come up with a worthy idea and solution that solves a problem we are currently facing ourselves.

What we learned

We learned project planning and management. In addition, we were able to learn the features of an efficient team, specifically goal clarity and role distinction.

What's next for Blinder Resume

We plan to complete, improve, and test our program. Following that, we plan to find partners with whom we can implement the program and study the changes. Based on these results we will continue to improve on the program and further develop it into a stage where it can fight inequality in resume screening.