WiNG.it

Inspiration

For this project, we were inspired by the struggles that young professionals face when practicing for interviews. Our team has experienced these struggles first-hand, and we wanted to create a resource that provides a more realistic way for users to practice for interviews. With that being said, many of the resources available for interview preparation do not simulate a realistic environment. For example, many interview preparation resources merely provide a list of questions to look over. Realistically in an interview, you will not have the luxury of reading a question and then crafting the answer with little pressure. Many helpful resources are also behind paywalls, which makes them inaccessible for some. WiNG.it is both more realistic and free, so that users can use it and become more confident in their interviews.

What it does

In every “interview round”, WiNG.it provides 5 behavioral interview questions to a user. WiNG.it’s mascot, Cornelius the Cat, reads these questions to the user. The user can then record their response via a microphone. With verbally delivered questions, we are able to mimic a more realistic interview experience. After the user completes all 5 questions, WiNG.it provides a transcript of each user response and the length of the response. The application also offers advice for answering each question as well as common traps that users may fall into when answering each question.

How we built it

We began this project by brainstorming ideas that we personally related to. As college students, all of us know that it is important to secure an internship. This commonality led to us talking about interview prep, and how it could be difficult to prepare for them without the right resources. Thus, we decided to make an application to help us prepare for interviews, specifically behavioral interviews, because of the particular lack of resources for practicing for these. To plan the structure of our application, we used a virtual collaborative whiteboard to jot down ideas and create some basic wireframes. Next, we researched libraries, packages, and datasets that we could apply towards our idea for WiNG.it. This helped us further hone in on our idea and decide on what was feasible for this hackathon. After solidifying our final ideas, we began finding data that we could use for the project. We found the data that we used on Kaggle, in the form of a PDF containing common behavioral interview questions and answers. At this point, we split up the work so that each team member had a focus. The members on the backend generated a CSV file from the PDF file using generative AI. The data in the file was accessed using the Pandas library on Python. These members also created a class and other data structures to represent questions and user responses, class functions, and a main function. On the frontend, members created HTML/CSS files for the home page, instructions page, interview question pages, and the results page. They first focused on getting the basic text and information down for these pages, then played around with the styling. Somewhere along the process, we decided that we wanted a custom mascot and background for the application. Thus, a member used Procreate to design Cornelius the Cat and his background. Other tasks that we completed were using Python functions with the route decorator to tell Flask what URL should trigger our functions. We also created functions to correspond with each of the HTML files we made previously.

Challenges we ran into

At the start of the hackathon, we had trouble pinpointing exactly what we wanted WiNG.it to be. We originally wanted to use machine learning algorithms and natural language processing in our project, but we could not find enough meaningful data to train a model. We also faced challenges with technical compatibility, as some teammates had difficulty configuring their setups to run our application smoothly. During the process of creating our application, we encountered problems with animation, incorporating audio, and connecting the back and front ends.

Accomplishments that we're proud of

This was each of the members’ first hackathon, so we are proud of pushing through the 36 hours. As for our project, we are proud of incorporating input and output audio. We did not think that these operations would be possible given the time constraints. We are also pleased with the end result. While it still has a way to go to be a fully-fledged interview preparation application, we are happy with the interface and the basic functionalities.

What we learned

Through creating the application, we became more familiar with accepting, parsing, and outputting audio in programs. We also became more familiar with the stages of developing an application. Lastly, we learned how to collaborate on a project to complete work efficiently.

What's next for WiNG.it

For the next iterations of WiNG.it, we would love to scrape the web for more behavioral interview questions and potentially train a machine learning model to detect the accuracy of a response, the tone of a response, or the quantity of filler words in a response. We would also like to incorporate verbal technical interview questions. Lastly, we would like to incorporate a way for users to create an account to keep track of their progress with WiNG.it.