Inspiration

Companies are constantly trying to improve the hiring process so that they can identify the employees that will be the most productive in their company. Oftentimes, it's difficult to differentiate between implicit biases and a candidate's performance. For example, a woman and man who are both outspoken about their opinions could be perceived in different ways; the woman could be considered rude or bossy, whereas the man could be seen as confident and idea-oriented. We want companies to be able to look at what happens during an interview objectively.

Within a company, the hiring process from start to finish takes many months. With the increasing number of applicants and the information that comes with it, the hiring process can get very tedious and lengthy. We wanted to make the entire hiring process more streamlined and data-driven.

What it does

With normaleyes, the guesswork and biases of interviews are removed, and information is displayed in a user-friendly, holistic evaluation. First, we record each interview and transcribe the audio using IBM Watson's Speech to Text. Next, we perform sentiment, emotion, and tone analysis on the conversation to identify the prevalence of the Big Five personality traits in a candidate: openness, conscientiousness, extraversion, agreeableness, and neuroticism. We also look at the candidate's sentiment towards their previous teams, and the emotional composition of each section of their interview.

Not only do we display the information we've synthesized on easy-to-read graphs, but we also support a live-time breakdown of the interview. Furthermore, we provide tools to facilitate other aspects of the hiring process, such as accepting a candidate and fulfilling travel reimbursements.

How we built it

We use a Nest Camera to record the audio and video components of interviews. We then use Bluemix APIs to analyze the interviews for sentiments, emotions, tones, and personality traits. Finally, the information is uploaded to AWS S3, where a React application allows users to visualize and understand the data.

Accomplishments that we're proud of

Our app is simple to use and easily integrates into any company. Moreover, it is scalable and has applications beyond corporate interviews. Not only can it provide feedback on a candidate from an interview, but it can also analyze any form of human conversation.

Share this project:
×

Updates