Inspiration

Interview season is here, and it's a tough stretch for many students. We wanted to streamline the preparation process and make it less intimidating for job seekers at any level.

What it does

Rehearsal Room is an AI-powered mock interview tool customized to the company you’re targeting. It reads each question aloud, records your on-camera responses, and then delivers a feedback summary with filler words, talking speed, STAR usage, and relevance.

How we built it

We built the interface in React and used the browser Media APIs to capture microphone and camera. ElevenLabs handles audio in both directions: text-to-speech reads each question to the user, and speech-to-text returns a transcript. We send that transcript to the Gemini API, which checks for filler words, evaluates whether the answer follows the STAR pattern, and judges how relevant it is to the question. Those results, along with basic timing data, are combined into a report that the user can review after each session.

Challenges we ran into

The biggest challenge we ran into was trying to integrate ElevenLabs into our project, as multiple aspects of the website relies on the speech-to-text and text-to-speech functions.

Accomplishments that we're proud of

We are super proud of how we were able to have ElevenLabs and Gemini work together. ElevenLabs was able to generate a transcript of the user's audio, and then it was sent to Gemini for analysis.

What we learned

We learned that clear data and simple rules matter most. With reliable audio transcriptions and a straightforward rubric (like STAR criteria), our feedback became sharper and more consistent.

What's next for Rehearsal Room

We plan to add social media features and more personalized coaching. A public leaderboard will showcase top-scoring mock interviews by company, so users can see strong examples and learn from peers. We’ll also let users upload a resume before the session, where the feedback will flag relevant experiences you didn’t mention, suggest specific tie-ins to the job, and propose stronger story angles.

Built With

Share this project:

Updates