QuackPrep

• 

  Logo

• 

  Home Page

• 

  Interview Page

• 

  Custom Questions Page

• 

  Job Portal

Inspiration

According to a survey, 92% of adults experience anxiety during job interviews. Often, interviewees fail to secure a job due to a lack of confidence, an inability to effectively showcase their skills, and insufficient practice. To address these issues, we created QuackPrep, a tool designed to help candidates better prepare for their interviews and secure their dream jobs.

What it does

QuackPrep simulates interview situations by asking questions and offering a video feature similar to real interviews. It captures both verbal responses and facial expressions. Later, it processes the data to provide feedback on how to deliver more effective responses and appear more agreeable. Additionally, it checks for grammar errors, repetitions, and stuttering, offering suggestions for clearer and more concise articulation.

We also offer curated questions from various companies and allow candidates to practice with their custom-made questions. Candidates can work with QuackPrep repeatedly until they can deliver the best answer, ultimately acing their interviews.

We also have a job board where students can search for and apply to different jobs.

How we built it

• Gemini to parse interview answers and improve response
• Google Cloud Speech-to-Text to transcribe audio from streaming user input using the microphone
• Google Cloud Text-to-Speech for asking interview questions
• face-api.js for facial emotion recognition
• socket.io to pass continuous streaming audio data from the backend to the frontend
• Procreate to design Mr. Goose
• Express - Node.js backend
• React frontend

Challenges we ran into

• Getting continuous streaming audio data from the backend to the frontend. We had to integrate socket.io for this feature.
• This was our first time using face recognition technology. We had to spend quite some time integrating face-api.js to identify emotional responses
• Getting Gemini to properly make suggestions for interview questions.

Accomplishments that we're proud of

We are proud of all the features we integrated in such a short time!

What we learned

• How to use face recognition technology
• How to integrate speech-to-text, text-to-speech, and Gemini seamlessly

What's next for QuackPrep

• Allow resume uploads and provide resume critiques
• Train QuackPrep with more interview questions for many other companies
• Work with recruiters to train custom LLMs for QuackPrep to make better suggestions
• Allow QuackPrep to detect different languages, to help with job prep for everyone around the world

Built With

• face-api.js
• gemini
• google-speech-to-text
• google-text-to-speech
• node.js
• procreate
• react.js
• socket.io
• tailwind.css