All Projects

CodePilot AI – Interview Preparation Platform

Open-source AI-powered coding interview preparation platform with 100 curated algorithm problems, an interactive syntax trainer across 9 languages, AI-guided hints with 3 coaching modes, AI code review, and a context-aware mentor chat. Full-stack Next.js + FastAPI architecture with Judge0 code execution.

Creator & Lead Developer2025 – Present
GitHub Live Demo

Stack

Next.js 14React 18TypeScriptTailwind CSSMonaco EditorFastAPIPythonOpenAI APIJudge0Docker

CodePilot AI – Interview Preparation Platform

Overview

CodePilot AI is an open-source, AI-powered coding interview preparation platform. It combines algorithm practice, interactive syntax training, and AI coaching into a single workspace. The platform supports 9 programming languages, 100 curated problems across 21 topic categories, and ~500 hands-on syntax exercises — all with integrated AI assistance that adapts to different learning styles.

What I Built

AI Coaching System

  • Three coaching modes: Strict (no hints, pure interview simulation), Interviewer (Socratic guiding questions), and Learning (direct explanations with pseudocode and snippets)
  • 5 progressive hint levels per problem — from subtle nudge to detailed walkthrough
  • AI code review: senior-engineer-level debrief after code execution with complexity analysis, edge case identification, and actionable improvements
  • Context-aware AI mentor chat with dual personality: coaching mode for problems (no spoilers) and tutoring mode for syntax (teaches freely with code examples)
  • OpenAI-compatible API — works with GPT, Claude, Llama, or any provider

Code Execution Engine

  • Judge0 integration for real-time code execution across 9 languages (Python, JavaScript, TypeScript, Java, C++, C, Go, Rust, C#)
  • Visible and hidden test case execution with execution time and memory stats
  • Structured test result reporting with pass/fail breakdown

Interactive Syntax Trainer

  • 13 curriculum categories with ~500 hands-on exercises across all 9 supported languages
  • Four-step exercise flow: explain → example → try → check
  • Progress tracking with AI feedback on exercise attempts

Full-Stack Architecture

  • Frontend: Next.js 14 App Router with Monaco Editor (same editor as VS Code), Tailwind CSS dark theme
  • Backend: FastAPI with 8 endpoints — code execution, hint generation, code review, syntax explanation and checking
  • API proxy: Next.js API routes proxy to FastAPI backend
  • Auto-save to localStorage with debounced persistence

Content System

  • 100 problem definitions as structured JSON (difficulty, topic tags, test cases, starter code per language)
  • 117 interactive trainer lessons organized by curriculum
  • 4 static reference lessons

Key Highlights

  • Full open-source AI coding platform from problem set to AI coaching to code execution
  • Multi-provider LLM support — swap between OpenAI, Anthropic, or local Ollama models via environment variable
  • Monaco Editor integration for production-grade code editing experience
  • Judge0 backend for sandboxed multi-language code execution