Salary data deep dives
Comparative questions answered from 8,049 published salary ranges at 197 companies · snapshot 2026-10-26
8,049postings with a published range
197companies
$200,000median midpoint
$149,978 – $253,000p25 – p75
The company pages answer "what does this company pay?".
These pages answer the questions that come after: is remote discounted, which function pays,
does a title predict a level, and what a location band does to the number.
Remote roles at these companies sit $200,000 at the median against office roles — but the gap flips for security and sales.
Data and ML leads, support trails. The spread between the top and bottom family is bigger than most promotion jumps.
"Senior" in a title is worth less than no adjective at all at a frontier lab. We show the labelling problem with numbers.
One city on a posting is a pay band. Three cities on a posting is a different job. The gap is large enough to be a data artefact worth knowing.
Most-covered companies
- openai — median $307,500 (629 postings)
- anthropic — median $345,000 (591 postings)
- databricks — median $216,104 (473 postings)
- elastic — median $138,800 (257 postings)
- snowflake — median $203,500 (239 postings)
- braze — median $121,000 (235 postings)
- oscar — median $157,010 (213 postings)
- brex — median $210,000 (200 postings)
- affirm — median $187,914 (196 postings)
- coinbase — median $200,078 (181 postings)
- ramp — median $214,300 (160 postings)
- airwallex — median $195,000 (157 postings)
Where the numbers come from
We read salary ranges that companies publish in their own job postings (Greenhouse,
Ashby, Lever, Workable), normalise them to annual USD, and take the midpoint of each
posted range. If a posting has no salary, it is not in the dataset. Sample size is
printed on every table because a median over 14 postings is not a median over 629.
Full method and the raw CSV: salary data.
Refresh log: changes.