🧠 Intelligent Interview Prep System

A sophisticated, multi-agent AI framework designed to prepare candidates for high-level technical interviews. The system orchestrates a team of specialized AI nodes using LangGraph, combining Hybrid Resume Parsing, Machine Learning Classifier Models, and Generative AI to deliver a bespoke, high-fidelity interview dashboard.

This Capstone & Hackathon edition consolidates both the FastAPI backend and a gorgeous, custom-styled React + Vite + Vanilla CSS dashboard into a unified project. The entire application runs concurrently with a single terminal command.

---

🚀 Key Highlights & Flagship Features

1. Cyclical Multi-Agent State Orchestration (LangGraph):

   • Orchestrates six specialized nodes (Analyst ➔ Strategist ➔ Planner ➔ Generator ➔ Validator ➔ Coach) working in a state-dependent feedback loop.
   • Enforces a quality control loop: if questions fail quality metrics, the Validator routes them back to the Generator for context-aware refinement.

2. Machine Learning Risk Detection & Calibrated Scoring:

   • Uses an ensemble classifier (Logistic Regression vs. Random Forest) calibrated using CalibratedClassifierCV to detect weak or unsubstantiated claims on a resume.
   • Computes a dynamic Readiness Score (0-100) mapping candidates to levels (Beginner, Intermediate, Expert) while penalizing vague and risky assertions.

3. Sleek Cyber-Dark Developer Dashboard:

   • A fully responsive React Single Page Application (SPA) designed using glassmorphism.
   • Features real-time, pulsing progress bars tracking agent steps, an interactive claim assessment table with glows, and tailored practice accordions with checklist points.

4. Zero CORS / Unified Deployment:

   • The compiled React client is hosted directly by the FastAPI web server. Running the backend immediately launches the frontend on the same port with zero cross-origin configuration required.

---

🏗️ Technical Architecture & Directory Layout

├── agents/             # Core LangGraph agent nodes & cyclical state graph
│   ├── resume_analyst.py
│   ├── question_strategist.py
│   ├── difficulty_planner.py
│   ├── question_generator.py
│   ├── validator.py
│   └── coach.py
├── ml/                 # Machine Learning modules (Risk Classifiers & Scorer models)
│   ├── risk_classifier.py
│   └── readiness_scorer.py
├── api/                # FastAPI endpoint controllers & static React web server
│   └── main.py
├── frontend/           # High-fidelity React + Vite + Vanilla CSS client application
│   ├── src/            # App.jsx states, index.css cyberpunk styles, assets
│   ├── index.html      # SEO metadata & Google Fonts imports
│   └── dist/           # Compiled, production-ready static assets
├── parsers/            # PDF and text parsers utilizing pdfplumber & regex
├── services/           # Utility integrations (Token tracking & cost counters)
└── utils/              # Config systems and logger helpers

More information:

-Live link: https://graphprep-interview-coach.onrender.com/

-Deep dive into the project: https://devpost.com/software/graphprep-multi-agent-ai-interview-preparation-system

---

🏃‍♂️ Quick Start (Single-Command Execution)

📋 Prerequisites

• Python 3.10+
• Node.js 20+ (only if rebuilding frontend files)
• Groq API Key (Get a free key at console.groq.com)

🛠️ Setup Steps

1. Clone the Repository:

   git clone https://github.com/Lokesh-get-git/capstone.git
   cd capstone

2. Install Python Dependencies:

   python -m venv venv
   # Activate virtual env:
   # Windows: .\venv\Scripts\activate
   # macOS/Linux: source venv/bin/activate
   
   pip install -r requirements.txt

3. Configure Environment variables: Create a .env file in the root directory and append your API key:

   GROQ_API_KEY=gsk_your_actual_groq_api_key
   # Optional: Tavily key for web searches during coaching
   TAVILY_API_KEY=tvly-your_key

4. Start the Application:

   uvicorn api.main:app --reload

That's it! Navigate to http://localhost:8000 in your browser to experience the dashboard.

---

🛠️ Developer Mode (Frontend Customizations)

If you wish to make live modifications to the React code and experience Hot Module Replacement (HMR):

1. Navigate to the frontend folder:

   cd frontend
   npm install
   npm run dev

   Your dev server will run on http://localhost:5173 and proxy backend requests automatically to port 8000.

2. Rebuilding production static files:

   npm run build

   This recompiles assets into frontend/dist, updating the unified FastAPI root immediately.