# Method

Source: our copy of 137,405,748 LinkedIn profiles (`blitz_people`), fetched in July and August 2026. One row is one profile; the row key is the profile URL. Each row has the current title, the company name and domain, and the city and country.

Match: a profile counts when its title contains "Member of Technical Staff", "Member of the Technical Staff", "Member Technical Staff", "Member of Tech Staff" or "Technical Staff Member", in any case, with or without a prefix such as Senior, Principal or Consulting. The regular expression is `member,?\s+(of\s+)?(the\s+)?tech(nical)?\.?\s+staff|technical\s+staff\s+member`, case insensitive.

Not counted: abbreviations (MTS, SMTS, LMTS, PMTS, AMTS, CMTS, DMTS) and misspellings such as "Member Techincal Staff". Abbreviations matter at a few companies: 2,326 more AMD profiles, 903 Salesforce, 460 GlobalFoundries, 323 NetApp, 320 Broadcom, 153 Nutanix, 110 Zoho, 90 Oracle and 7 OpenAI profiles use one. With them AMD has 5,174 and Zoho 5,090.

Employer: the company domain on the profile. VMware profiles sit under broadcom.com. Rows with no domain are in the country and city counts but not in the employer ranking.

Two passes, which agree: a full scan of every row for the ranking, countries and cities (`employers.csv`, `countries.csv`, `cities.csv`), and a count through the company domain index for 83 named domains, which also gives each company's total profiles and title ladder (`companies.csv`).

Ladder: the word before "Member" decides the rung (Junior and Associate, Senior or Sr, Lead, Principal, Consulting, Distinguished); no word means the plain title.

Limits: these are profiles, not headcounts. People write their own titles, many keep an old title after a promotion, and our copy covers some companies better than others. 1,786 of the 7,346 OpenAI profiles and 846 of the 22,787 Zoho profiles have no title at all. Company totals include some people who are not employees, such as investors, forum members and ambassadors who list the company.

Script: `tools/report_mts_facts.py` in the ProfessionLens repository.