interview.ai | Devpost

Inspiration

We were motivated to create this after the difficulty we personally, and our friends, faced while getting tech internships.

What it does

It has a whiteboard area that converts handwritten code to actual text and allows the interviewee to run it in the browser! We also have a video/audio recording feature for the interviewee to practice talking to an actual interviewer by responding to randomly generated technical interview questions. Additionally, we allow the interviewee to download their session for review later. Moreover, we graph the user's statistics such as sentiment, confidence, and concept frequency and use machine learning techniques to predict future values.

How we built it

We used MyScript for whiteboard, React for all the front-end web components, Node.js and Python in Google App Engine for the back-end, Tensorflow/Keras/Scikit-learn/Numpy for the machine learning, and plotly for the graphs.

Challenges we ran into

Connecting and synchronizing all the various components was very challenging as well as using GAE for the first time and using ML techniques (i.e. LR, NF, etc) in a practical project.