Interview Maestro

This is Us: UI Welcome Page
Your Personal Interview Assistant: Personalization Interaction Page
What Makes Us Special: We serve a diverse age range, including middle school students.
What Makes You Special: We Accept Multimedia Forms Of Personalization
Our Options: We Provide Both Quick Audio and Video Interview Options
Your Personal Assistant: Quick Practice With AI Assessment
Your Personal Assistant: Live Interview With AI Assessment
Your Personal Assistant: Live Interview With AI Assessment
Sample Live Interview With Gemini Feedback
Sample Live Interview With Gemini Feedback

Interview Maestro is an AI-assisted interview practice web app designed to help users rehearse answers and improve delivery across three tracks: Academic, Social, and Career. After each practice interaction, the app returns actionable feedback with an overall score and component-level signals such as tone, pacing, and clarity. It also includes a real-time live interview experience with a video-interview-style interface.

Interview prep is often stressful, and the stress itself degrades performance: people speak too fast, lose structure, and sound uncertain even when the content is strong. The goal of this project was to make practice feel low-friction and repeatable: practice → receive feedback → iterate. The Monet-inspired visual theme supports that goal by aiming for a calmer, less intimidating practice environment.

We learned:

How to turn subjective communication quality into clearer, actionable dimensions (e.g., tone, pacing, clarity), rather than generic feedback.
How to structure model responses so that the frontend can reliably parse and render them (productizing generative output).
Practical frontend-backend integration patterns for an AI workflow (React UI ↔ Flask API ↔ Gemini).
Browser constraints for a live, interview-like experience (permissions, user experience, and integration via a dedicated live module).

Challenge:

Consistency of AI output: Generative models can vary across runs, so the prompts and response format must be constrained enough for stable UI rendering.
Latency and UX: The end-to-end loop (submit → model call → response) requires careful UI state handling so users understand what the app is doing.
Live interview constraints: Camera/mic permissions and browser differences can introduce friction, so the live experience needs robust handling and clear user guidance.

The interactive demo is deployed based on a forked code repo that is forked from the main repo for testing. Interactive demo.