Overview

LeetCoach is an innovative AI-powered platform developed during a hackathon to transform coding interview preparation. Designed to assist students and professionals, it integrates real-time guidance, voice-enabled AI support, and secure code execution, offering an engaging and effective solution for mastering technical interviews, particularly with LeetCode problems.

The Challenge

Preparing for coding interviews poses significant obstacles for many students:

Lack of Direction: Beginners often struggle to identify a starting point or structure their practice effectively.
Isolation: Solo practice can feel demotivating without immediate feedback or interaction.
Limited Feedback: Standard solutions lack personalized insights into code quality and optimization.
Accessibility Gaps: Traditional tools rarely accommodate diverse learning preferences, such as auditory learners.

The Solution

LeetCoach addresses these pain points with a robust set of features:

Practice Problems with Code Execution: Write and run code directly within the platform.
Voice-Enabled AI Support: Hands-free guidance via an interactive AI assistant.
Real-Time Feedback: Instant insights on code quality, complexity, and edge cases.
Natural Voice Responses: Powered by Neuphonic’s text-to-speech for an immersive experience.
Progress Tracking: Built-in timer to monitor and enhance problem-solving speed.

Technical Architecture

LeetCoach leverages a modern tech stack for performance and scalability:

Frontend: Next.js, TypeScript, and Tailwind CSS for a responsive, user-friendly interface.
Backend: Python AIOHTTP service ("codedriver") for secure code execution.
AI Integration: Gemini API via Google AI client for intelligent, real-time feedback.
Voice Technology: Web Speech API for transcription, paired with Neuphonic API for natural responses.
Code Editor: CodeMirror for an intuitive coding environment.
Security: Sandboxed execution environment with restricted imports to ensure safety.

Key Features Implementation

Voice Interaction

A standout feature is the voice-enabled AI assistant, implemented as follows:

const setupVoiceInteraction = () => {
  const recognition = new webkitSpeechRecognition();
  recognition.onresult = async (event) => {
    const transcript = event.results[0][0].transcript;
    const response = await getAIResponse(transcript);
    speak(response);
  };
  recognition.start();
};

const speak = async (text) => {
  const audioUrl = await neuphonicAPI.textToSpeech(text);
  const audio = new Audio(audioUrl);
  audio.play();
};

typescript

This enables users to interact hands-free, receiving natural-sounding feedback.

User Experience Design

Intuitive Layout: Clean design with easy navigation to problems, editor, and AI assistant.
Accessibility: Voice support enhances usability for diverse learners.
Motivational Tools: Timers and feedback scores encourage continuous improvement.

Challenges and Learnings

The development process presented valuable lessons:

Secure Code Execution: Built a sandboxed environment to safely run user code.
Prompt Engineering: Refined AI prompts for concise, actionable feedback.
API Integration: Overcame documentation challenges with Neuphonic and cloud services.

These hurdles deepened our understanding of secure coding, AI optimization, and third-party integrations.

Future Enhancements

Expanded Problem Set: Additional questions across varying difficulty levels.
Language Support: Extend beyond Python to other programming languages.
Collaborative Mode: Enable pair programming functionality.
Advanced AI Coaching: Tailored guidance for specific interview strategies.
Interview Simulation: Timed challenges mimicking real interview scenarios.

Conclusion

LeetCoach redefines coding interview preparation by blending AI-driven insights with voice technology and a secure, user-centric design. Developed during a hackathon, it showcases innovative problem-solving and technical expertise, making it a standout addition to my portfolio as a Computer Science student at the University of Manchester.

LeetCoach