AI Interview Coach

Hone your interview skills at time until perfection!

Comment

Inspiration

Job interviews are stressful, especially technical ones, and many candidates struggle to prepare effectively on their own. We wanted to create an accessible tool that provides personalized interview practice with real-time feedback, helping candidates improve their skills and confidence before facing real interviews.

What it does

AI Interview Coach is an application that simulates real interview experiences for different roles (Software Engineer, Product Manager, Data Scientist). It uses speech recognition to capture responses, delivers questions via text-to-speech, provides immediate constructive feedback after each answer, and generates a comprehensive report analyzing overall performance, strengths, areas for improvement, and actionable recommendations.

How we built it

We created a serverless architecture using:

  • Frontend: React with TypeScript and Vite
  • Backend: AWS Lambda functions, API Gateway, and Bedrock
  • Speech functionality: Browser Speech Recognition API and AWS Polly
  • AI analysis: Claude 3.7 Sonnet via AWS Bedrock for generating personalized feedback and reports
  • Additional analysis: AWS Comprehend for sentiment analysis

The application follows a three-stage process: interview setup, interactive question-answer session, and comprehensive report generation. We used AWS Serverless Application Model (SAM) for infrastructure as code deployment.

Challenges we ran into

  • Getting API call to work (CORS)
  • It doesn't work, pretty big issue

Accomplishments that we're proud of

Tried:

  • Building a fully functional end-to-end interview simulation experience
  • Leveraging Claude 3.7 Sonnet to provide highly personalized, role-specific feedback
  • Implementing speech recognition and synthesis for a more immersive experience
  • Creating a serverless architecture that's scalable and cost-effective
  • Designing an intuitive interface that guides users through the entire interview process

What we learned

  • Effective prompt engineering techniques for instructing LLMs to provide structured feedback
  • Best practices for implementing speech recognition in web applications
  • Building serverless applications with AWS Bedrock and Lambda
  • Creating responsive React components that handle asynchronous processes gracefully
  • Integrating multiple AWS services into a cohesive application

What's next for AI Interview Coach

  • Making the API calls work
Share this project:

Updates