View the project LIVE and see the instructions in the video as well as our Github.

Purpose

Cloak is an applications platform that removes any personal biases from the hiring process, by hiding an applicant's personal information fields, as outlined by the form creator, until after they have been accepted. Fields the form creator could choose to "make secret" could be name, gender, race, etc, and will be visibly differentiated. We also provide the form creator with various other tools to help in their decision making process, such as an NLP-powered summarizer, and a JSON API for an employer to run their own processing on.

Inspiration

Now you may be wondering what inspired us to make Cloak. So let's break it down. At first, one of our team members watched a video explaining the inherent biases that people have and how it affects everything in their daily lives, from employment to shopping. This reminded of the blind auditions that music companies use with great effect. And this is where we realized... if the music industry can adopt this practice, why can't anyone else.

So with that, and the massively increasing unemployment caused by COVID-19, Cloak came into existance and now it's fully functional on our website cloak.theavid.dev so let's dive in.

Development

The core functionality of Cloak resides in the Django open source Python web framework. It allows us to create flexible models, stored in a PostgreSQL database, without writing a single line of SQL. Through Django's admin pages, Cloak hides all personally identifiable information to remove potential bias from applications. Cloak is hosted on a dedicated server running an Nginx reverse proxy. Our fabulous domain name is provided by Namecheap with DigitalOcean acting as a nameserver and host. All traffic is secured through LetsEncrypt SSL certificates ensuring that your private data is safe.

Challenges

Of the three of our members, each one of us was using a different OS and browser, between us covering Chrome, Firefox, Safari on Windows, Arch Linux, and MacOS. This led to some differences in the front-end, and meant we had to make sure all of our commits worked on all three browsers.

However, this also meant that the final product was as polished as possible, working on a large number of systems.

Accomplishments

The biggest accomplishment that we are proud of is the model system that integrated perfectly with the PostgreSQL database and stored all of the data. It worked seamlessly right off the bat because we took a proper amount of time thinking about the best implementation. Getting the site completely functional and into productions was also very rewarding since it means that everyone can try it out and see what it does.

Learning

Some of us were new to Django, and got the chance to familiarize with it through this project. We also got to experiment with things such as TextRank's summarization and secure random number generators. We took an inside look at how other providers make their forms and came up with a solution that expanded the possibilities.

Future

We would continue to expand on the functionality for form creators, allowing a file upload option (where we would remove instances of name, gender, race, and other private information), expand our export options to include .txt, .pdf, .md, etc., and allow for custom email acceptance/rejection messages.

Another possibility is making the secret answers completely secure using end to end encryption so they never leave the user's browser. This will also require some changes to the backend infrastructure because of how the data is currently stored, but it's a goal we'd love to accomplish.

Share this project:

Updates