Tech Pay Index

Home › Methodology

Methodology

How the figures on this site are built, and what they can and cannot tell you.

Source

Every day we read the public job boards of 249 AI and technology employers through the documented public endpoints of three applicant-tracking systems: Greenhouse, Lever and Ashby. These are the same endpoints that power each employer's careers page. No login is used and no personal data is collected.

What counts as pay

Only postings that publish a pay range are used; today that is 20,637 postings. Where the job board supplies pay as structured data, that is used. Otherwise the range is read from the posting text. Hourly rates are multiplied by 2,080, weekly by 52 and monthly by 12. Each posting contributes the midpoint of its annualised range.

When a posting lists several pay zones, the zone matching the posting's location is used. If none matches, the middle zone is used, never simply the highest.

Sales roles that publish on-target earnings are kept apart from base salary and are not included in the figures on this site.

Counting rules

What these figures are not

History and trends

Daily snapshots have been kept since 2026-11-10. Change-over-time figures will be added once there are at least 60 days of history and both ends of a comparison have 20 or more postings.

Role and level labels

Role families and levels are assigned from job titles by fixed rules. Titles with no level word are labelled "Level not stated" and are not assumed to be mid-level.

Data updated 2026-11-22. Advertised base pay for open roles, excluding equity and bonuses. Midpoint of the annualized pay range published in each job posting (public Greenhouse, Lever and Ashby boards). Hourly ×2080, monthly ×12. Duplicate postings (same company, title, location and range) counted once. Sales roles that publish on-target earnings (OTE) are reported separately from base salary.

Full methodology and limits