In a time when face to face conversation is becoming less prominent, there are few resources available for students to practice interviews for prospective job and internship opportunities. Most college students have a computer or a phone and every college student could use some practice with interviews. We set out to find a way every student has easy access to a critical job-finding resource like interview practice.

What it does

Taylr, true to its name, tailors uses' interview skills to help you get better at interviews and help you understand your mistakes and improve on them.

Enter some quick questions about the type of interview you'd like to test out, your name and phone number and get a call from Taylr, our interviewer. Taylr will ask you a set of questions which you will answer in real time. Following the interview, Taylr will send you a link to view a dashboard in a simple, sleek and responsive UI to view in-depth analytics about your speech and sentiment.

How we built it

We built our web application in Python Flask which serves a form that users can use to input their information. Once the form is submitted we used Twilio's API to call the number and start the phone interview. The user then responds to every question asked by Taylr, who takes all the responses and checks for filler words, grammatical errors, sentiment, and tone of speech. Lastly, Taylr sends a message with the URL to view the results.

Challenges we ran into

In terms of user experience, we had some challenges in figuring out the user flow to contain as few steps as possible. Large amounts of data (in graphs or numbers) can be overwhelming to users, so we made it a priority to create a dashboard that conveys as much information as possible to the users but also display it in a simple interface.

We wanted to implement machine learning to determine whether or not the users response adequately answered the question that Taylr asked. We explored using a neural net and a vector representation of words but we were not able to find a corpus of training data.

Accomplishments that we're proud of

We were able to create an entire platform with a streamlined user experience. Each of us worked on learning new skills and used this opportunity to create something with meaningful.

What's next for Taylr

Moving forward we hope to add the ability for users to have multiple interviews with Taylr and have the ability to compare results and track their progress overtime. Also stepping away from IBM Watson and implementing out own natural language processing algorithms. We would want to expand our platform to individuals across many career fields by providing a large database of technical questions specific to a wide variety of positions. In order to achieve this, we would plan to crowd source interview questions in order to build our database. Another potential feature would be for users to have the option to play back audio clips of their interview so that they could further analyze their responses.

+ 4 more
Share this project: