EquiHire

Inspiration

As women in STEM, we have experienced and witnessed bias and discrimination in hiring and workplace environments. We wanted to create a tool that promotes fairness and equity by addressing hidden biases in job postings and recruitment processes.

What it does

EquiHire uses AI to analyze job descriptions for biased language, providing inclusive job recommendations making the hiring process fairer for all.

How we built it

We began by integrating a pre-trained language model from Hugging Face, specifically designed for bias detection in text. The model was capable of flagging gendered terms, coded language, and other subtle biases commonly found in job descriptions. We compiled an open-source dataset of job listings that included job titles and descriptions from multiple sectors. We developed a custom bias evaluation algorithm that assigns a score to each job posting. The score considers the presence of gender-coded words and occurrence of racially coded or exclusionary phrases. The platform’s frontend was built using React, while the backend runs on Flask.

Challenges we ran into

We ran into many challenges, one being connecting our frontend and our backend. We worked with both flask and react, and it was difficult to take elements from our backend and display them on our frontend. We solved this problem by finding online resources such as documentation and videos that explain both flask and react in a more cohesive way. By going through these resources, we determined proper ways to use both the framework and library together.

Accomplishments that we're proud of/What we learned

As it's most of our teams first hackathon, working towards a project in a short amount of time was challenging at first. However, we worked together and persevered towards creating a project we're proud of. It was really interesting to learn how to use specific libraries and frameworks, especially to create a working website.