Inspiration
I've seen the issue of interviews all around me in real life, so I wanted to create a simple interview practice tool that would make preparations less intimidating and boost confidence.
What it does
I built InterviewBuddy as a website using React and TypeScript. The user chooses an interview role, receives AI-generated interview questions, and answers each question either using text or speech recognition. After completing the interview, the answers are analyzed and the user gets instant personalized feedback.
How I built it
I first brainstormed my ideas and gathered them into Figma and Bolt to have an idea of the prototype and the aspects that I wasn't sure how to make. Then, I learned how to connect a web app to an AI API, worked with API responses, used React to manage an interactive interview, and used the browser's Speech Recognition API for voice input. I improved the visuals of the website using Coolors color themes, Figma/Bolt gradient colors, and Tailwind CSS, while also using Git and Github to create a repo.
Challenges I ran into
One of my biggest challenges was getting the AI API and speech recognition to work correctly. I had to debug the API authentication because I first used Meta Llama, which is gated and would not work when I tested. I also had to make sure the user's answers were properly saved for each question.
Accomplishments that I'm proud of
I'm pretty proud of what I've made overall. The color theme and looks of the website is modern and I got a lot of inspiration while researching online. The transitions and gradients are smooth as well, and the speech works as intended.
What I learned
I learned how to connect a website to an AI API using FeatherLess AI, and how to work with AI-generated responses to display them. I also learned how to implement speech recognition and how to manage user answers between questions. Debugging was extremely useful as many parts of the code were bugging along the way.
What's next for InterviewBuddy
I'm looking to add more career categories and interview questions, improve the AI feedback, make the speech recognition work more reliably, add some statistics for users to track over multiple interviews, and improve the UI so it can be more accessible on mobile devices.
Built With
- featherless
- github
- lucide-react
- qwen
- react
- tailwind
- typescript
- vite
- web-speech-api
