Pay evidence retains its source vintage · employer signals refreshed 23 September 2026

Public market evidence brief

Role families and source-backed comparisons

A concise view of five benchmark families, with original currencies, RWF reference displays, evidence grades, and direct source links.

✓Evidence rule
No amount appears without a direct source URL.
Compensation rows50public, source-linked observations
Official documents12employer ATS or public-sector records
Hourly / task rates8kept hourly, not monthlyized
Employer records176screened public comparator records

Evidence assistant

Ask from the linked records

The assistant only summarizes rows that have a source link in this workspace.

One-screen management brief

The role spans five market families

These public comparators are arranged by the work they benchmark. Nairobi figures stay visibly Nairobi figures; RWF is shown only as a cross-currency display. No median is invented and no source is treated as payroll unless the source says so.

Read this firstScope, evidence grade, original currency, and direct source are shown together.

Role-value conclusion

Combined scope should not be assessed against an annotation-specialist IC benchmark alone.

The comparison set covers annotation and perception quality, people and throughput management, technical data workflows, cross-functional delivery, and engineering responsibility.

One-page comparison table

Published range bands by work family

Direct links below
Benchmark familyWhat the comparator coversModest publishedMiddle publishedHigh publishedEvidenceSource
How to read the bandsModest is the source's low observation, middle is its midpoint where published, and high is its high observation. Each cell keeps the source currency and adds an RWF reference display; no blended median is created.

Evidence grades

AEmployer-published compensation or official public-sector document.
BLarge employee-reported sample, retained as a comparator rather than payroll.
CSmall employee-reported or aggregated sample; directional only.
DJob-board expectation, historical listing, platform rate, or secondary source.

Role-scope evidence

International sources confirm the discipline is technical

Official global postings connect human-data operations with annotation quality, tooling, metrics, workforce planning, data pipelines, and engineering delivery. Their compensation is not used as a Rwanda benchmark.

Evidence discipline

What the table keeps separate

  • Base salary vs total compensationLevels.fyi is shown as total compensation, not base.
  • Monthly vs annual vs hourlyUnits remain in the source form; no task rate is monthlyized.
  • Local vs regionalKenya is a regional comparator, not a Kigali equivalence.
  • Current vs historicalExpired listings stay visible only as dated evidence.

Coverage lanes now tracked

Employer coverage is broader than the five benchmark rows

Telecom / NOC / drive-testRURA, MTN, Airtel, Ericsson, Nokia, KTRN
Government / digital systemsRDB, RURA, RSSB, C4IR, RISA
Health / data infrastructureCHAI, PIH, UGHE, digital health programmes
Energy / field engineeringREG, EUCL, EDCL, KivuWatt, Bboxx
AI / cloud-workDigital Umuganda, TELUS, Welo, Wing, DataAnnotation
Pan-African / international remoteAGL, ALX, Andela, Remote, Careers in Africa

Employer signals, kept separate

Benefits and career signals are not compensation evidence

These records support targeting and preparation only. Salary bands remain in the comparison table; employee-reported, secondary, historical, and unavailable signals stay labelled.

Employer / sectorPublic signalWhat it supportsVerifiedEvidence / warningSource

Linked evidence table

Compensation observations

49 rows

Portable records

Download the public tables

These files contain only public comparator evidence. The original source URL is retained in every money row.