PG-AGI Technical Interviewer 🤖💼

An autonomous, agentic technical screening system designed to conduct deep-dive technical interviews. Built with a focus on RAG (Retrieval-Augmented Generation) and NVIDIA NIM, this system evaluates candidates based on their resumes and foundational knowledge from core ML textbooks.

🌟 Key Features

• Adaptive Interview Pipeline: Dynamic difficulty gating that adjusts based on candidate performance.
• RAG-Powered Evaluation: Contextualizes questions using a knowledge base built from industry-standard textbooks (e.g., Tom Mitchell's Machine Learning).
• RAG Triad Observability: Real-time metrics for Context Relevance, Answer Faithfulness, and Answer Relevance.
• Glassmorphism UI: A premium, modern interface with smooth animations and interactive chat.
• Session Persistence: Resume interviews exactly where you left off.
• Comprehensive Dashboard: Detailed post-interview summary with performance analytics and confidence scoring.

🏗️ Architecture

Backend (Python/FastAPI)

• Engine: Orchestrates the interview flow and state management.
• Knowledge Base: ChromaDB-powered vector store for semantic retrieval of technical concepts.
• RAG Pipeline: Integrates NVIDIA NIM API for high-performance inference and evaluation.
• Resume Parser: Extracts key skills and experience to tailor the interview.

Frontend (Next.js/TypeScript)

• Framework: Next.js 15 with React 19.
• Styling: Tailwind CSS 4.0 for a sleek, responsive design.
• Animations: Framer Motion for premium transitions and micro-interactions.
• State Management: React Context/Hooks for real-time interview synchronization.

🚀 Getting Started

Prerequisites

• Python 3.10+
• Node.js 18+
• NVIDIA NIM API Key

Backend Setup

1. Navigate to the backend directory:

   cd backend

2. Create a virtual environment and install dependencies:

   python -m venv venv
   source venv/bin/activate  # Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Configure environment variables in a .env file:

   NVIDIA_API_KEY=your_api_key_here
   DATABASE_URL=sqlite+aiosqlite:///./pgagi.db

4. Ingest textbooks into the knowledge base:

   python scripts/ingest_textbooks.py --sample

5. Start the server:

   uvicorn main:app --reload

Frontend Setup

1. Navigate to the frontend directory:

   cd frontend

2. Install dependencies:

   npm install

3. Start the development server:

   npm run dev

📊 RAG Triad Observability

The system employs the RAG Triad to ensure the quality of the AI's reasoning:

1. Context Relevance: Is the retrieved knowledge relevant to the question?
2. Faithfulness: Is the answer derived strictly from the retrieved context?
3. Answer Relevance: Does the answer directly address the candidate's query or the role requirements?

📜 License

Distributed under the MIT License. See LICENSE for more information.

---

Built with ❤️ by ThePriyanuj