WOICE

Meeting page
Title page
Proposed Architecture

Inspiration

Many women around the world feel that they have an unfair disadvantage or bias against them while giving interviews. There are multiple studies and first hand experiences we've heard and read that motivated us to think and implement of a solution that eliminates this unfortunate factor from interviews.

What it does

Our project is an anonymous audio conferencing tool that will modulate voice and output a more gender neutral voice that will make it difficult to identify the gender of the candidate giving interviews.

How we built it

We initially started with an architecture diagram and split the tasks amongst ourselves and worked parallely. We used various technologies like React, Flask, Pytorch and Tensorflow to implement our project

Challenges we ran into

We faced multiple challenges while implementing this solution. One of them was using webrtc to connect in real time to pierce or sending of voice modulated audio. Also, our deep learning model for authenticating users takes a bit of time to validate the user. Integrating everything together.

Accomplishments that we're proud of

We're proud of the progress we've made in a short time and that we were brave enough to take on a complicated task and project that will solve a large problem. Built a image recognition model using Pytorch.

What we learned

We learned that hackathons are challenging and a great platform for learning a lot quickly. We all learned various tools like Pytorch React, and integrating other technologies like Domain.com, Google clous to enhance our project

What's next for WOICE

We would like to integrate voice analysis so that the interviewer can also understand the emotions and various other characteristics of the candidate. Also, we want to leverage the cloud and Google AutoML effectively for computer vision.