Insterview

Breakdown Page

Inspiration

Video interviews are rapidly becoming an integral part of the recruitment process, but there is a lack of tools that let prospective candidates better prepare themselves for what is usually the first round of the recruitment process. This is also stressful for those who haven't encountered video interviews before, such as first year undergraduates. We see the potential to develop a tool that bridges this gap in the market.

What it does

Our interview simulator aims to measure:

your sentiment positive or negative
your eye contact,
engagement from vocal tonality,
body language
confidence from tonality - whether you are building rapport or breaking rapport based on tonal direction
how much you hesitate (um,er,hm)
am I speaking too quickly
how you sound on your microphone on camera especially to the employer as a recording

How we built it

Our existing implementation is built as a web app on a JavaScript, with Node and Express as the back-end and the Material Design framework for the front-end. We implemented a set of APIs to interface between our front-end web application and the Cognitive Service APIs.

The Cognitive Service APIs from Microsoft Azure play an integral part of our web application in the following ways:

Video Indexer API: used to analyse the video, including extracting the script, the sentiment and keywords.
Text Analytics API: used to extract keywords from the answers the interviewee given
Face API: used to analyse the facial expression and body language of the interviewee

Challenges we ran into

The challenges during the hack were mostly technical - most of us were not familiar with developing web applications so it took time to learn the basic groundwork of developing an Express application. There were also issues with integrating the Cognitive Service APIs within the existing JavaScript stack and handling responses.

Accomplishments that we're proud of

Our biggest accomplishment is being to strike a good balance between engineering the technologies behind the product and devising a viable business plan during the short 9 hours.

What we learned

A Lot. From the technical skills to the merging between programming, design and business ideas, we have learned how to integrate the different disciplines into a complete product.