InterviewInsider

• 

  Logo

• 

  Homepage

• 

  Lifecycle

• 

  Motivation

Inspiration

As job and internship seekers, our group understands the importance of making a solid impression during interviews. However, knowing how to improve can be difficult without access to practice partners, unbiased feedback, or access to a pool of questions. Wouldn't it be fantastic if you had an on-demand partner that could provide you with real-time feedback on your responses?

What it does

InterviewInsider generates interview questions for you and records your answers to provide you feedback based on the content of your answer, your vocal tone, and the speed of your speech.

How we built it

We obtain the user's verbal input and send its transcription to Gemini, Google's large language model, for content feedback. We determine users' speech speed per minute and tone using natural language processing and pre-trained machine-learning models. This feedback is presented to users on a simple yet straightforward webpage where they can reflect on their responses and make changes.

Challenges we ran into

The main challenge was integrating these different technologies through a Flask server and identifying how to properly assess the tone and inflection of the user's voice.

Accomplishments that we're proud of

We're proud that our web app can use Google Gemini's API for accurate user feedback and determining their vocal tone using open-source machine learning models. We are also proud of creating a fully functional website that seamlessly uploads and downloads audio files.

What we learned

We learned how API calls to large language models, open-source machine learning models, and natural language processing libraries work.

What's next for InterviewInsider

InterviewInsider aims to create a conversational AI that asks regular questions and follows up to clarify users' answers. Moreover, we would like to fit the experience based on the role and industry the user is applying for.

Built With

• flask
• fullstack
• gemini
• javascript
• llm
• machine-learning
• natural-language-processing
• python
• webapp