Jason Chen

About This Site

How Jason Chen's portfolio website is built — from raw data to deployed pages.

Work Experience

Technical Supervisor

Wistron

2025-07 – Present

Responsible for full-stack software development and evaluating new technologies to drive digital transformation across Wistron's multinational enterprise operations. Also writes technical tutorials and builds internal template/scaffolding modules to support knowledge sharing within the team.

Senior Software Engineer

Wistron

2024-07 – 2025-06

Responsible for full-stack software development and evaluating new technologies to drive digital transformation across Wistron's multinational enterprise operations. Also writes technical tutorials and builds internal template/scaffolding modules to support knowledge sharing within the team.

Software Engineer

Wistron

2021-07 – 2024-06

Responsible for full-stack software development and evaluating new technologies to drive digital transformation across Wistron's multinational enterprise operations. Also writes technical tutorials and builds internal template/scaffolding modules to support knowledge sharing within the team.

Software Engineer

Cihong Technology Co., Ltd.

2019-04 – 2021-06

Developed and maintained software across platforms, including dual-platform apps (Android, iOS), frontend web apps, admin dashboards, backend API services, and automation scripts.

  • Managed the company's internal firewall, VMs, and AWS cloud servers
  • Participated in project planning, proposing process improvements from a technical perspective
  • Met with clients to address technical questions and estimate development timelines based on requirements

Education

National Taiwan University

B.S., Geological and Earth Sciences/Geosciences

2014-09 – 2018-06

Certifications

Claude Code in Action · Anthropic · 2026-06
Certificate of completion: Introduction to Model Context Protocol · Anthropic · 2026-06
Certificate of completion: Introduction to subagents · Anthropic · 2026-06
Certificate of completion: Introduction to Claude Cowork · Anthropic · 2026-06
Certificate of completion: Introduction to agent skills · Anthropic · 2026-06
Certificate of completion: Claude code 101 · Anthropic · 2026-06
Certificate of completion: Claude 101 · Anthropic · 2026-06
Prompt Design in Vertex AI Skill Badge · Google · 2024-07

Data Collection

All project data originates from CLAUDE.md files that live inside each of 115+ local project repositories. A custom Claude Code skill (/generate-tech-profile) reads each CLAUDE.md, extracts structured metadata, and outputs bilingual JSON files under tech-profile/. The scan is triggered manually via /scan-projects, which walks ~/project/ and appends new entries to a master projects-list.md index.

Data Format

The tech-profile/ directory holds a flat set of JSON files, all bilingual (English + Traditional Chinese):

  • meta.json — profile summary, years of experience, total project count

  • domains-{en|zh}.json — 9 skill domains with per-skill proficiency levels (expert / proficient / familiar)

  • product-groups-{en|zh}.json — 23 business product groups, each listing constituent project IDs and roles

  • projects/{ID}.{en|zh}.json — per-project detail: tags, category, status, description, domain

  • tag-index-{en|zh}.json — flattened index of 1,976+ unique technology tags for fast lookup

Presentation

The website is a Next.js 15 static export (output: 'export') deployed to Firebase Hosting. All data is loaded via static JSON imports at build time — no server-side fetch or database. Joy UI's CssVarsProvider handles dark/light mode without JavaScript overhead. Language switching (EN ↔ ZH-TW) uses a lightweight React context backed by localStorage.

Tech Stack

Next.js 15
TypeScript
Joy UI (MUI)
Firebase Hosting
Claude Code CLI
pnpm

SEO & AI Discoverability

Every page exports server-rendered metadata (title, description, OpenGraph). Product detail pages include JSON-LD structured data (schema.org/SoftwareApplication) so search engines and AI crawlers can understand each product's context. A /llms.txt file follows the emerging LLM-crawler convention for direct AI agent consumption.

/llms.txt · /robots.txt