Career Objective

Independent developer. A 2021 graduate with a history degree in Hong Kong. Studying computer science at HKMU full-time. After graduation, I find my real passion for frontend deveploment and throw all my energies into pursuing related knowledge. Proficient in using popular frontend tech stacks. Hoping to join a tech company where I can put my skills to good use.

Core Skill

• HTML, CSS, JavaScript
• React.js
• Next.js
• TypeScript
• GraphQL
• Node.js
• SQL
• Git

Portfolio

• 1. Personal Website

  Developed a personal website using the Next.js framework

  • Responsive Design: Modern and responsive layout optimized for various devices.
  • Blog Section: User-friendly blogging platform with SSG and ISR for fast page generation and real-time updates.
  • AI-Powered Summaries: Integrated Deepseek Chat AI's API to automatically generate concise and informative summaries for blog posts, enhancing reader engagement and content accessibility.
  • Content Management: Admin panel for managing blog posts, pages, and media using Strapi.
  Next.js Framer Motion Strapi GraphQL

  Website

  https://jackycheung.dev

• 2. DoraSearch — Manga Full-Text Search

  A Doraemon manga full-text search engine that went semi-viral on Social Media (7,000+ likes & saves), built with a React SPA frontend and Cloudflare Workers + D1 backend.

  • Community-Validated Side Project: Launched as a personal learning project and shared on Xiaohongshu (小紅書), where it received 7,000+ likes and saves plus 300+ positive comments from users and developers.
  • AI OCR Pipeline at Scale: Processed ~8,000 manga pages (2GB+) with Qwen3-VL, extracting structured text with high OCR accuracy.
  • Cost-Aware Query Design: Designed a hybrid pagination strategy (cursor-based for prev/next, offset for jump-to-page) plus Cloudflare Cache to minimize scanned rows.
  • React Search Experience: Implemented a single-page React app with TanStack Router and TanStack Query for URL-driven search state, client-side caching, and smooth navigation across result pages.
  React TanStack Query Cloudflare Workers SQLite FTS Python Qwen3-VL

  Website

  https://dorasearch.cc/

• 3. PastScan — Document OCR & Search Platform for Researchers

  A high-performance SaaS platform for historical researchers to upload, automatically OCR, and semantically search scanned PDF documents.

  • AI OCR Pipeline: Advanced text recognition powered by Gemini 3.1
  • Metadata Extraction: Automated extraction of structured metadata (Table of Contents, title, author, date) for research citation generation
  • Semantic Search: Vectorization of OCR results using pgvector to enable highly accurate semantic search
  Next.js PyMuPDF Supabase Cloudflare Workers Gemini 3.1 pgvector

  Website

  https://histori-ai.pastscan.workers.dev

Languages

• Cantonese (Native)
• Mandarin (Proficient)
• English (Good command)