Global Salary Data by Skill & Role (620k jobs)

Key findings

1. 620,752 job postings with stated pay from 137,004 employers in 55 countries, January 2024 to June 2026, every figure annualised and converted to US$. 56% are in the United States; 22 countries have at least 30 postings and 14 have at least 400.
2. Country medians span 24× among the 14 countries with 400+ postings: Switzerland $117,470, United States $60,000, Ireland $55,319, Austria $53,933 at the top; India $7,558, Brazil $4,896 at the bottom.
3. Software engineering pays a median $120,805 in the US, $78,947 in the UK and $9,012 in India at market exchange rates, a 13× US–India gap.
4. 109,469 postings carry skill tags. Among skills with 200+ US postings the best paid are ai ($134,000), c++ ($131,000), aws ($128,000), kubernetes ($125,750).
5. In India the best-paid tagged skills (200+ postings) are ca ($31,977), llm ($31,977), microservices architecture ($31,395), bfsi sales ($29,070), against a national median of $7,558.
6. The UK sample is small (2,129 postings, median $46,581): few UK postings in the collection state pay, so UK figures carry wide uncertainty. 1,553 postings with absurd or unconvertible pay were dropped before any statistic was computed.

Salary percentiles by country (US$ per year)

Countries with at least 10 postings. "Posted in" is the most common posting currency (Gulf and some Asian listings were advertised in Indian rupees by an Indian board). Countries under 100 postings are indicative only.

Role family: United States vs United Kingdom vs India vs rest of world (median US$)

Median annual pay in US$; a median is shown only where the region has 20+ postings for the role. Sorted by the US median.

Seniority: United States vs United Kingdom vs India vs rest of world (median US$)

Seniority from the posting or title, else from years of experience asked.

Industry / occupation group: US vs UK vs India vs rest of world (median US$)

Top 50 most requested skills with salary percentiles: US vs India vs rest of world

Skills as tagged in postings (lower-cased). Skill tags exist mainly on US, Indian and Singapore postings; UK postings in this collection carry no skill tags. A median is shown where the group has 20+ postings.

Top 50 highest-paying skills in the United States (150+ postings, 20+ employers)

Ranked by median annual pay among postings that tag the skill (US$ at the stated FX rate).

Top 50 highest-paying skills in India (100+ postings, 20+ employers)

Ranked by median annual pay among postings that tag the skill (US$ at the stated FX rate).

Salary percentiles by country and role family (countries with 400+ postings)

Groups with 30+ postings, sorted by country size then median.

Postings and median pay by month and region

Collection is uneven over time; monthly counts reflect collection as well as hiring. Medians shown for 20+ postings.

Exchange rates used (approximate 2025 averages)

Every local figure is divided by "units per US$". Rates are fixed, so differences between countries also reflect exchange rates, not purchasing power.

Full dataset on request

This release publishes about 100,000 postings (a fixed set of whole companies). The complete set of 620,752 postings in 55 countries, with the same columns, is available on request for research, journalism and other fair use.

• Email [email protected] with your name, organisation and a short description of how you will use the data, or
• use the Fliar contact page.

Column dictionary

18 columns in the sample CSV.

Method

1. Scope. Postings with a title, an employer, a country and stated pay, published 2024-01-01 to 2026-06-30.
2. Deduplication. Postings with the same employer, normalised title and month are one posting; the copy with the most structured fields is kept.
3. Annualisation. Hourly × 2,080, daily × 260, weekly × 52, bi-weekly × 26, semi-monthly × 24, monthly × 12. Indian figures between 1 and 60 in an annual field are lakhs per annum; Indian figures between ₹1,000 and ₹99,999 in an annual field are monthly pay. Singapore listings from one board that carried monthly pay multiplied by 144 are divided by 12. Postings without a currency take the country's own currency.
4. Currency conversion. One fixed table of approximate 2025 average market rates (units per US$): AED 3.6725, AUD 1.55, BRL 5.6, CAD 1.4, CHF 0.83, CLP 950, COP 4100, CZK 22.5, EUR 0.89, GBP 0.76, HKD 7.8, HUF 360, IDR 16400, ILS 3.6, INR 86, JPY 150, MXN 19.3, MYR 4.35, NZD 1.72, PEN 3.7, PHP 57.5, PLN 3.8, QAR 3.64, RON 4.45, SAR 3.75, SEK 10, SGD 1.31, THB 33.5, TRY 39, UAH 41, USD 1, UYU 40, VND 25500, ZAR 18. The full table, with posting counts, is printed above.
5. Absurd values dropped, and how many. Annual pay below US$15,000 (US), US$10,000 (other high-income countries), ₹30,000 (India) or US$1,200 (elsewhere); above US$1,000,000 (US), ₹2.5 crore (India) or US$600,000 (elsewhere); weekly travel-contract healthcare pay annualised with the wrong unit; implausible hourly, weekly or low-wage-title figures; postings whose currency differs from the country's own (for example US jobs posted in rupees); Indian overseas-placement and job-scam postings. Counts: currency differs from country 538, below floor 372, weekly travel-contract pay mislabelled 309, implausible for role 80, intern pay above ceiling 80, above ceiling 73, implausible hourly rate 55, ambiguous rupee figure 24, currency not convertible 19, hourly/daily rupee pay 2, implausible weekly rate 1; Indian overseas postings 482; Indian job-scam postings 210.
6. Labels. role_family from the posting, else keyword rules on the (mostly English) title, else the category. seniority from the posting or title, else from years asked. region_group puts every country other than the US, UK and India in "Rest of world".

7. Privacy and provenance. URLs, source names, descriptions and recruiter details are not published; posting IDs are hashes. Postings whose employer or title names a job board or data provider are dropped.

8. Sample. Each company key is ranked by sha1(key). Keys are taken in rank order until the file holds max(25%, 5,000) rows (capped at 100,000).

9. Profile links. Links are built from Fliar's company index and checked against it. A row whose company has no public profile has an empty link.

Limitations

• Coverage is very uneven by country: the United States, Canada, France, India and Germany hold most postings; 33 countries have fewer than 30 postings and the UK only about 2,100. Country figures under a few hundred postings are indicative.
• The queue idea expected 47 countries with skill-level detail. Skill tags exist mainly on US, Indian and Singapore postings, so skill percentiles describe those markets; elsewhere the dataset supports role-level percentiles only.
• Fixed exchange rates convert currencies but do not adjust for purchasing power or for exchange-rate moves over 2024–2026.
• Posted pay is an offer, not realised pay; it usually excludes bonus, equity and benefits. Hourly pay is annualised at 2,080 hours.
• Role keyword rules are English; non-English titles (France, Germany) fall more often into "other".
• Collection is uneven over time, with most rows in 2026; month-to-month changes mix collection with hiring.

Cite this dataset

@misc{fliarbi_skill_salary_percentiles_global_2026,
  author    = {{Fliar BI}},
  title     = {{Global Salary Data by Skill & Role (620k jobs)}},
  year      = {2026},
  publisher = {Fliar BI},
  url       = {https://fliarbi.com/datasets/skill-salary-percentiles-global/},
  note      = {Sample release. Licence CC BY-NC 4.0}
}