Iris- Your AI Interviewer(Audio+ Visual+ Live Cam+ Feedback)🎙️📹✨

This is a submission for the Google AI Studio Multimodal Challenge

💡 What I Built

Iris is an AI-powered Interview Coach designed to help candidates practice, improve, and excel in interviews.
Instead of reading generic interview guides, Iris creates a personalized interview experience by combining resume understanding, live camera interaction, and structured feedback — powered by Google AI Studio (Gemini multimodal models).

It feels just like a face-to-face mock interview with a professional hiring manager, but available anytime, anywhere.

🌟 Iris Features

• 📄 Resume Scan – Upload PDF/image, extracts key skills, experience, education & generates a concise summary.
• 🎤 Live AI Interview – Face-to-face AI interviewer with camera & mic, adapting questions to your resume & responses.
• 🤖 Personalized Feedback & Coaching – Gets strengths, areas to improve, and STAR method evaluation from your session.
• 💾 Download Feedback Report – Save detailed feedback as a document.
• 🔗 Share Progress – Share reports with friends/mentors for collaborative review.
• 📜 Interview History – Timeline of past interviews with scores & feedback summaries.
• 📂 Previous Transcripts – View transcripts with questions & answers to track progress over multiple sessions.

🎥 Demo

👉 Video

Screenshots

App Source Code : Link

🧑‍💻 Features in Detail
1. 📄 Resume Scan + AI Summary

Upload your resume as a PDF or image.

Iris uses Gemini multimodal capabilities to extract:

• ✅ Key skills
• ✅ Experience highlights
• ✅ Education & certifications
• Generates a concise summary to guide the interview.

2. 🎤 Live AI Interview (Audio + Visual + Camera)

• Iris acts as a virtual interviewer.
• You interact face-to-face using your camera and microphone.
• The AI adapts its questions to your resume and responses.
• Covers both behavioral (STAR method) and technical questions.

3. 🤖 Personalized Feedback & Coaching

After the session, Iris analyzes the entire transcript.

• Feedback includes:
• 🌟 Strengths (clarity, confidence, relevance)
• 📌 *Areas to Improve *(with actionable advice)
• 🎯 STAR Method Evaluation (Situation, Task, Action, Result)
• Output is structured, easy to understand, and encouraging.

4. 📑 Save & Share Your Progress

💾 Download Feedback Report as a document.

🔗 Share with friends/mentors for collaborative review.

📜 Include Interview History

• Show a timeline of past interviews:
• Date & time of each interview
• Overall score or feedback summary

📂 Include Previous Interview Transcripts

• Provide transcripts or summaries of past interviews
• Questions asked
• Candidate answers Helps track progress across multiple practice sessions.

🛠️ How I Used Google AI Studio

Iris is powered by Gemini multimodalfeatures:

• Resume Parsing (File + Text Input)
• Uses gemini-2.5-flash to extract structured insights from resumes.
• Interview Simulation (Chat API)
• Creates a dynamic chat session where Gemini acts as a hiring manager.
• Feedback Analysis (JSON Schema)
• Ensures structured responses:

{ "strengths": ["Good clarity", "Relevant examples"], "areasForImprovement": ["Expand on technical depth", "Use STAR format"] } 

📦 Gemini Services Used

1. gemini-2.5-flash ⚡

Why used:

• Fastest multimodal Gemini model.
• Handles resume parsing from PDFs/images quickly.
• Great for interactive chat-based interviews where low latency is critical.

import { GoogleGenAI, Chat, Type } from "@google/genai"; ✅ generateContent() → Resume extraction + feedback generation ✅ chats.create() → Interactive interview simulation ✅ JSON schema → Well-structured coaching insights 

🚀 Why Iris?

• 🎯 Problem: Many candidates struggle with interview preparation — either they don’t know what to expect or lack constructive feedback.

• 🌟 Solution: Iris provides a safe, intelligent, and supportive environment to practice, with AI that:

• Understands your resume

• Conducts live interviews

• Gives actionable coaching feedback

• Helps you improve over time

• With Iris, you no longer have to walk into an interview unprepared.