shiganluo/skill/career-research

Career Research: Chinese Company Career Portals

Research real job postings on Chinese company career portals. Output verified positions with direct links.

Workflow

1. Go directly to company career pages — do NOT use delegate_task (times out on heavy SPA sites) or DuckDuckGo curl (unreliable for Chinese queries). Use browser_navigate directly.
2. Use `browser_console` to extract structured data — NOT browser_snapshot (truncates at ~80 elements). Run JS to extract job titles, IDs, locations, descriptions, and detail-page URLs from the DOM.
3. Try multiple search keywords — "生物信息" may return 0 results while "医疗" returns 268. Always try at least 3 variants.
4. Verify each position — click through to detail page or extract real href attributes. Never fabricate URLs.
5. Check graduation period — Chinese campus recruitment explicitly states 毕业时间 (graduation window). Verify the target graduation date falls within range.

Company Career Page Directory

Tech Companies (校园招聘 / Campus Recruitment)

Biotech / Genomics Companies

Platform Notes

• Beisen (北森) sites (*.zhiye.com): Job listings are SPA modals, no direct detail URL. Use the search page URL as reference. Extract job IDs (JXXXXX) via browser_console.
• ByteDance: Best URL structure. Detail pages accessible via direct URL. ByteIntern positions explicitly state graduation period.
• Tencent: 毕业时间 in homepage banner. 应届生 vs 实习生 have different windows.

browser_console Extraction Patterns

Extract job links from ByteDance

(() => {
  const links = document.querySelectorAll('a[href*="detail"]');
  return JSON.stringify(Array.from(links).map(a => ({
    text: a.innerText.substring(0, 150),
    href: a.href
  })).filter(r => r.text.length > 5));
})()

Extract full job text from any SPA

(() => {
  const main = document.querySelector('main') || document.body;
  return main.innerText.substring(0, 5000);
})()

Find Beisen job links

(() => {
  const allLinks = document.querySelectorAll('a');
  return JSON.stringify(Array.from(allLinks)
    .filter(a => a.innerText.includes('J'))
    .map(a => ({text: a.innerText.substring(0, 100), href: a.href})));
})()

Pitfalls

1. delegate_task with browser toolset times out on heavy SPA career sites (600s timeout, 25-46 API calls wasted). Always use browser tools directly.
2. DuckDuckGo HTML search (html.duckduckgo.com/html/?q=...) returns empty for most Chinese job-related queries. Do not rely on it.
3. browser_snapshot truncates at ~80 elements, hiding job listings. Use browser_console with innerText extraction instead.
4. URL-based search doesn't work on Meituan (/web/campus/search?keyword=XXX → 404). Must use in-page search input.
5. BGI careers page at /careers/ returns 404. Real portal is at genomics.zhiye.com (found via footer "加入我们" link href).
6. Alibaba defaults to English (lang=en). Must append ?lang=zh for Chinese.
7. Campus recruitment timing: 2027届秋招 typically opens August-September 2026. Some companies (ByteDance) open early; others (Tencent, Alibaba) may still show 2026 positions in early August.

Search Keyword Strategy

For bioinformatics/medical AI positions, try these keywords in order:

1. "生物信息" — most specific, may return 0 on tech company sites
2. "医疗" — broader, returns many results on ByteDance (268 positions)
3. "基因" — for genomics-specific roles
4. "算法" — for algorithm engineer positions (very broad)
5. "健康" — catches health-tech positions

Resume Building

For creating or improving resumes from existing materials and wiki content, see references/resume-building-workflow.md.

See Also

• references/company-findings-2026-08.md — Detailed findings per company from August 2026 session
• references/resume-building-workflow.md — Resume extraction, gap analysis, and improvement workflow