Interview.io

• 

• 

• 

• GIF

• 

Inspiration:

As second year students preparing for co-op, one of our biggest fears was messing up our interview. We wanted to create a website that goes through an individual's resume and asks interview level questions, some personalized according to the individuals resume. The best part? Interview.io will give you a sophisticated analysis of how you did!

What it does:

A user can enter the role he wants and describe it slightly. On the next page, the user would submit a pdf of his resume. We can go through this resume and with the help of chat gpt we make 6 interview level questions for the individual. The individual can record themselves and answer the questions and this recording is then graded by AI on 4 different criteria all tailored to the job their applying too. The individual can look at their score and know what they should be working on to ace their interview.

How we built it:

We worked on our front end using HTML, CSS and JavaScript. Using a google API, we were able to convert the speech to text in python. In another python program, using OpenAi's GPT 3.5 API, we were able to provide questions and evaluate the answers given by the individual. We made a program in HTML, CSS and JavaScript which asks for permission to use the mic and camera and allows them to record their answers.

Challenges we ran into:

We were having trouble in making the speech to text python program. We tried importing speech_recognition. We also tried to install chocolatey and ffmpeg but after installing it, we faced a lot of issues and could not continue. We also tried using Open AI's whisper but we were not able to make them work despite numerous efforts. We finally ended up using the google API. 3 of our members were completely new to html, css, javascript, and Figma. In fact, this was the first time any of us used any of the frameworks/languages used in Interview.io (with the exception of Python).

Accomplishments that we're proud of:

This is the first website we had ever made and while it was an ambitious first web dev project we were able to get it working. We are really proud of the fact that we were able to make our backend work and view it on the website. It did not seem like an easy task in the beginning but as a team, we all did it together! We are also proud of the proompt engineering we were able to do with the OpenAI API to make the interview questions. The figma designs we made for our website was another proud accomplishment since this was our first introduction/attempt at Web Design and UI/UX.

What we learned:

We were all just beginning programmers, having learnt HTML and CSS a couple of days ago. Using what we learned and studying about more complex functions made us feel more comfortable using these languages. We used JavaScript too and learned it while googling different functions and implementing them in our program.

What's next for Interview.io:

We plan on studying JavaScript in detail and improving our website so that we can make it available in the public domain. We also want to add a camera recording feature to analyze eye contact and facial expressions throughout the interview. Also adding a functionality to create an account and save progress on your interview stats to view the change in your stats over multiple attempts is something we are interested in. Adding a feature to save your final interview results as an infographic PDF is another feature we brainstormed.

Built With

• css
• express.js
• figma
• flexbox
• html
• javascript
• node.js
• openai
• python