E-Learning Platform Builder

Build a React TypeScript video player component for an e-learning platform. Requirements: 1) Uses HTML5 video element (no external player library), 2) Custom controls: play/pause, seek bar, volume, playback speed (0.75x, 1x, 1.25x, 1.5x, 2x), fullscreen, 3) Progress tracking: saves current position to backend every 10 seconds via POST /api/progress with { lessonId, userId, position, completed }, 4) Resume playback: on mount, fetches last position from GET /api/progress/:lessonId and seeks to it, 5) Completion detection: fires onComplete callback when 90% of video watched, 6) Keyboard shortcuts: space (play/pause), arrow keys (seek 10s), 7) TypeScript with proper types. Include all state management with useState/useRef.

Write a Node.js Express endpoint POST /api/certificates/generate that creates a PDF certificate of completion. Requirements: 1) Uses PDFKit (npm install pdfkit), 2) Certificate design: landscape A4, dark navy background (#0A1628), gold border, centered layout, 3) Content: "Certificate of Completion" heading, student full name (large, serif font), course title, completion date, instructor name and signature line, platform logo placeholder, unique certificate ID (UUID), 4) Saves PDF to /tmp/certificates/{certificateId}.pdf, 5) Returns { certificateId, downloadUrl } in response, 6) Separate GET /api/certificates/:id endpoint serves the PDF file, 7) TypeScript. Include full error handling.

Build a Node.js course recommendation API endpoint GET /api/recommendations/:userId. Requirements: 1) Collaborative filtering: find users with similar enrollment history, recommend courses they completed that the target user hasn't started, 2) Content-based filtering: recommend courses sharing tags with the user's completed courses, 3) Combine both signals: 60% collaborative, 40% content-based, 4) Return top 6 recommendations with { courseId, title, reason, score }, 5) reason should be human-readable: "Students who completed X also loved this" or "Matches your interest in [TAG]", 6) Cache results in Redis for 1 hour (use ioredis), 7) Fallback: if insufficient data, return top-rated courses in user's enrolled categories. TypeScript + Drizzle ORM.