CodeIntelligent AI — Production AI Coding Interview Platform

A production-grade, enterprise-ready AI Coding Interview Platform engineered with Next.js (App Router), Monaco Editor, MongoDB Atlas, Mongoose, and a multi-language sandboxed code execution engine paired with conversational AI evaluation.

Key Architecture & Capabilities

graph TD
    User["Candidate / Admin"]
    Frontend["Next.js 15 Client (React 19, Monaco IDE, Tailwind, Lucide)"]
    API["Next.js Route Handlers & Server Actions"]
    Auth["JWT Session Cookies + Argon2/Bcrypt + RBAC"]
    DB[("MongoDB Atlas (Mongoose Models)")]
    AI["AI Provider Abstraction (Google Gemini 3.7 Flash / Fallback)"]
    Runner["Multi-Language Sandboxed Code Runner"]

    User --> Frontend
    Frontend --> API
    API --> Auth
    API --> DB
    API --> AI
    API --> Runner
    Runner -->|Test Results, Peak Memory, Wall-Clock Time| API
    AI -->|Real-Time Dialogue, Hints & 6-Dimension Scorecard| API

Features Matrix

1. Authentication & Role-Based Access (RBAC)

Candidate & Admin Roles: Separate access permissions for taking interviews vs. administering problem banks.
Secure Password Hashing: Salting and hashing via bcryptjs / Argon2.
Stateless & Resilient JWT Sessions: Signed using jose with HS256 and stored in secure HTTP-only cookies.
1-Click Demo Login: Instantly switch between sample Candidate and Admin profiles for rapid testing.

2. Live Conversational AI Interviewer

Powered by Google Gemini 3.7 Flash: High-speed reasoning with structured JSON generation.
Adaptive Socratic Probing: Interactively challenges candidates on edge cases, memory allocations, and Big-O complexity.
3-Tiered Progressive Hint System:
- Level 1: Conceptual nudge on data structure choice.
- Level 2: Algorithmic approach (two-pointer, DP state, sliding window).
- Level 3: Concrete invariant and edge-case pointers.
Voice Support: Integrated speech recognition and text-to-speech audio playback.

3. Professional Monaco Code IDE

Multi-Language Support:
- Python 3
- JavaScript (Node.js)
- TypeScript
- Java 17
- C++ 20
- Go 1.22
IDE Features: Syntax highlighting, theme toggle (vs-dark / light), font resizing, starter code reset, custom test case injector.
Keyboard Shortcuts: Ctrl+Enter / Cmd+Enter to trigger instant test runs.

4. Sandboxed Code Execution Engine

Resource Limiting: CPU limits, memory caps (128MB), and execution timeout killers (preventing infinite loops).
Dual Mode: Containerized Docker execution + native isolated subprocess runner.
Comprehensive Test Results: Standard output stream (stdout), error stream (stderr), wall-clock execution time in milliseconds, and peak memory in KB.

5. Multi-Dimensional Rubric Scoring & Evaluation

Test Correctness (0-100): Evaluated against visible and hidden test suites.
Code Quality & Cleanliness (0-100): Readability, modularity, idiomatic patterns, naming conventions.
Algorithmic Efficiency (0-100): Detects actual Time and Space Big-O complexity vs. optimal theoretical baseline.
Edge Case Coverage (0-100): Assesses handling of nulls, empty collections, negative numbers, and boundary limits.
Problem Solving Approach (0-100): Structured decomposition and speed.
Technical Communication (0-100): Articulation during interview conversation.
Actionable Growth Roadmap: Personalized suggestions, recommended topics, and optimal reference solution diffs.

6. Candidate & Admin Dashboards

Candidate Dashboard: Skill proficiency matrices, score trends over time, completed session history, and recommended study paths.
Admin Portal: Problem bank CRUD, test case authoring, and platform health telemetry.

Project Structure

ai-coding-interview-platform/
├── src/
│   ├── app/
│   │   ├── (auth)/login/page.tsx         # Login with 1-click demo switcher
│   │   ├── (auth)/register/page.tsx      # Registration with role preferences
│   │   ├── admin/page.tsx                # Admin challenge manager & audit logs
│   │   ├── dashboard/page.tsx            # Candidate analytics & history
│   │   ├── interview/
│   │   │   ├── setup/page.tsx            # Interactive interview wizard
│   │   │   ├── [id]/page.tsx             # 3-Pane Live IDE + AI Interview workspace
│   │   │   └── report/[id]/page.tsx      # Multi-dimensional evaluation report
│   │   ├── problems/page.tsx             # Curated problem bank
│   │   ├── layout.tsx                    # Root layout with navbar & theme
│   │   ├── page.tsx                      # Landing page with interactive preview
│   │   └── api/                          # REST API endpoints (Auth, Interview, Code, Problems)
│   ├── components/
│   │   ├── ui/                           # Button, Card, Badge, Modal, Tabs, Progress
│   │   ├── ide/                          # Monaco editor, Test panel, Console output
│   │   ├── interview/                    # AI chat box, Hint modal, Timer bar
│   │   ├── dashboard/                    # Score analytics, History table, Recommendations
│   │   └── navbar.tsx                    # Top navigation header
│   └── lib/
│       ├── db/
│       │   ├── mongodb.ts                # Resilient cached Mongoose connection
│       │   └── models/                   # User, Problem, Interview, Submission, Evaluation
│       ├── auth/                         # Password hashing, JWT signing, Cookies, RBAC
│       ├── ai/                           # Gemini 3.7 Flash provider & evaluation rubric
│       ├── runner/                       # Sandbox dispatcher & language harnesses
│       └── validations/                  # Zod input schemas
├── scripts/
│   └── seed-problems.ts                  # Curated problem database seeder
├── tests/
│   ├── code-runner.test.ts               # Sandbox execution & timeout tests
│   ├── auth.test.ts                      # JWT session and password tests
│   └── evaluator.test.ts                 # AI evaluation rubric tests
├── Dockerfile                            # Production multi-stage Docker build
├── docker-compose.yml                    # App + MongoDB + Redis orchestration
├── package.json
└── tsconfig.json

Getting Started

Prerequisites

Node.js: v20+ or v22 LTS
MongoDB: Local MongoDB instance (mongodb://localhost:27017) or MongoDB Atlas URI
Python: (Optional for local Python runner fallback) Python 3.10+
Docker: (Optional for full containerized deployment)

1. Installation

git clone <repo-url>
cd ai-coding-interview-platform
npm install

2. Configure Environment Variables

Copy .env.example to .env.local:

cp .env.example .env.local

Fill in your configuration:

MONGODB_URI=mongodb://localhost:27017/ai_interview_platform
JWT_SECRET=super-secret-key-change-this-in-production-min-32-chars
GEMINI_API_KEY=your_gemini_api_key_here
NEXT_PUBLIC_APP_URL=http://localhost:3000

(Note: If GEMINI_API_KEY is not provided, the platform automatically activates its intelligent fallback engine for continuous offline/local development.)

3. Seed Problem Bank & Demo Users

npm run seed

This populates standard DSA problems (Two Sum, Longest Substring, Coin Change, Maximum Subarray, Valid Parentheses) and sets up demo Candidate (candidate@example.com) and Admin (admin@example.com) accounts with password password123.

4. Start Development Server

npm run dev

Open http://localhost:3000 in your browser.

Testing

Run the automated test suite covering code sandboxes, timeouts, JWT authentication, and AI evaluation rubrics:

npm test

Docker Deployment

To launch the full production stack (Next.js App + MongoDB + Redis) in Docker:

docker compose up --build -d

Access the application at http://localhost:3000.

Security Best Practices

Never Executes on App Thread: User-submitted code runs strictly in isolated sub-sandboxes with enforced execution timeouts and memory ceilings.
Zero API Key Leakage: AI credentials and database URIs remain strictly server-side.
HTTP-Only Cookies: JWT authentication tokens are protected from cross-site scripting (XSS).
Zod Validations: All incoming API requests and AI JSON responses are strictly validated before execution.

License

MIT License. Built for real-world technical interviewing preparation.

Name		Name	Last commit message	Last commit date
public		public
scripts		scripts
src		src
tests		tests
.env.example		.env.example
.gitignore		.gitignore
Dockerfile		Dockerfile
README.md		README.md
docker-compose.yml		docker-compose.yml
next-env.d.ts		next-env.d.ts
next.config.mjs		next.config.mjs
package-lock.json		package-lock.json
package.json		package.json
postcss.config.mjs		postcss.config.mjs
tailwind.config.ts		tailwind.config.ts
tsconfig.json		tsconfig.json
vitest.config.ts		vitest.config.ts

Navigation Menu

Repository navigation

About

Repository files navigation

CodeIntelligent AI — Production AI Coding Interview Platform

Key Architecture & Capabilities

Features Matrix

1. Authentication & Role-Based Access (RBAC)

2. Live Conversational AI Interviewer

3. Professional Monaco Code IDE

4. Sandboxed Code Execution Engine

5. Multi-Dimensional Rubric Scoring & Evaluation

6. Candidate & Admin Dashboards

Project Structure

Getting Started

Prerequisites

1. Installation

2. Configure Environment Variables

3. Seed Problem Bank & Demo Users

4. Start Development Server

Testing

Docker Deployment

Security Best Practices

License

About

Releases

Packages

Contributors

Languages

Latest commit

History

Folders and files

Repository files navigation

CodeIntelligent AI — Production AI Coding Interview Platform

Key Architecture & Capabilities

Features Matrix

1. Authentication & Role-Based Access (RBAC)

2. Live Conversational AI Interviewer

3. Professional Monaco Code IDE

4. Sandboxed Code Execution Engine

5. Multi-Dimensional Rubric Scoring & Evaluation

6. Candidate & Admin Dashboards

Project Structure

Getting Started

Prerequisites

1. Installation

2. Configure Environment Variables

3. Seed Problem Bank & Demo Users

4. Start Development Server

Testing

Docker Deployment

Security Best Practices

License

About

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages