InterviewX

InterviewX is a technical interview platform designed for real-time 1-on-1 coding assessments. It combines a synchronized IDE, video conferencing, and an automated code judging engine to simulate a professional interview environment.

Technical Stack & Versions

• Frontend: React.js (v18.3.1) for UI components and state management.
• Styling: Tailwind CSS (v4.0) with DaisyUI for a high-performance, utility-first design system.
• Backend: Node.js and Express.js for RESTful API architecture.
• Real-time Video: Stream Video SDK (v1.24.0) for low-latency communication.
• Real-time Chat: Stream Chat SDK (v9.23.0) and React components (v13.9.0) for synchronized messaging.
• Database: MongoDB for flexible session and user data storage.
• Authentication: Clerk for secure user management.
• Code Execution: Piston API for isolated, multi-language execution.
• State Management: TanStack Query (v5.0) for efficient data fetching and caching.
• Code Editor: Monaco Editor for a VSCode-like development experience in the browser.

Key Features

• VSCode-Integrated Editor: Uses the Monaco Editor API to provide syntax highlighting and a professional coding experience.
• Multi-Language Code Execution: Connects to a stateless execution engine (Piston API) to compile and run code in C++, JavaScript, Python, and Java.
• Real-time Video & Chat: Integrated Stream Video SDK with support for screen sharing, camera/mic toggling, and instant messaging.
• Automated Test Runner: A custom logic layer that injects test cases into user code and validates results using string normalization.
• Strict Session Guard: Backend logic that limits room capacity to exactly 2 participants (Host + 1 Candidate) to ensure interview integrity.
• Live Dashboard: Progress tracking with success rates and detailed history of previous coding sessions.

Technical Implementation

Code Judging Pipeline

The core of the platform is the "Judge" logic. Instead of just displaying output, the system:

1. Wraps user-submitted code with a test-case driver.
2. Sends the payload to an isolated execution environment via the Piston API.
3. Sanitizes the raw stdout using Regular Expressions to remove formatting noise and trailing whitespaces.
4. Returns a "Success" or "Fail" verdict based on a comparison with hidden expected outputs.

Session Management & Security

To ensure session security, I implemented a locking mechanism in the Node.js/Express backend. When a user attempts to join a session, the server validates the current participant count. If the count is $\ge 2$ and the user is not the designated host, access is denied.

To solve the issue of "ghost" participants (where a user leaves but the count remains high), I utilized Stream webhooks to listen for call.session_participant_left events, triggering an automatic decrement of the participant count in the MongoDB database.

Installation

1. Clone the repository: git clone https://github.com/vikassalgude/InterviewX.git
2. Install dependencies: npm run build (This installs both frontend and backend dependencies).
3. Set up environment variables for Clerk, Stream, and MongoDB.
4. Start the production server: npm start.