Deemwar Monads & Abstractions

AI hiring, counted

We counted 20,971 tech job posts. Here is what the AI hiring numbers show.

20,971 postings
185 company boards
13 months of HN
9.0% AI-titled

Deemwar data note, 12 November 2026. Every figure is our own count from free public sources; the rules we wrote before counting, the method notes and the results are published with it.

We counted 20,971 tech job posts: four numbers

We went in with a view: AI and forward-deployed roles are booming, and small companies are the ones missing out. Before counting, we wrote down four questions and what result would prove us wrong. Then we counted. Parts of our view held. One part did not.

What we counted

Company job boards20,971 postings on the public job boards (Greenhouse, Lever, Ashby) of 185 companies, on 12 November 2026. We chose the companies; we tried 321 and 136 had no public board in those three systems. This is a sample of well-known tech employers, not the job market.
Hacker NewsThe monthly “Who is hiring?” thread, 13 months from November 2025 to November 2026. This is the only source where we have more than one point in time.
RulesFixed word rules applied to job titles. “AI-titled” means the title has AI, ML, LLM, GenAI, machine learning or agent in it. “Forward-deployed” means the title says so.

1. The base rate: about 1 post in 11 has an AI term in its title

Share of all postings by title type

Of 20,971 postings, 1,879 (9.0%) are AI-titled and 749 (3.6%) are AI or ML engineer titles. 383 (1.8%) are forward-deployed. Software-titled roles are 6,542 postings (31.2%); within those, 17.5% are AI-titled and 4.6% forward-deployed.

One label we expected to see barely exists: titles that combine AI and automation number 7 out of 20,971.

2. Over 13 months on Hacker News: forward-deployed up, AI-engineer roles flat

Hacker News hiring posts, 13 months

Posts naming forward-deployed rose from 5 of 391 (1.3%) in November 2025 to 10 of 228 (4.4%) in November 2026. Forward-deployed peaked at 5.2% (16 of 307) in June 2026 and has eased since. Posts mentioning an AI or ML engineer role went from 19.4% to 18.4%, which is flat by our pre-set two-endpoint rule; month to month the share ranged from 14.0% to 22.4%. Posts mentioning any AI term went from 58.8% to 63.6%.

The number of posts per month fell from 391 to 228. The shares above are of a shrinking pool.

For comparison, Indeed Hiring Lab reports US software postings up almost 15% since February 2025 while all postings fell 7%, with 71% of the net software growth between May 2025 and May 2026 in senior roles (Indeed Hiring Lab). Lightcast counts forward-deployed postings up more than 1,000% from January to August 2026 against the same months of 2025 (Fortune). Our Hacker News series points the same way on forward-deployed, from a much smaller base.

3. Forward-deployed roles sit at a handful of companies

Five companies post 65.3% of forward-deployed roles

Of the 383 forward-deployed postings, 250 (65.3%) are at five companies: Databricks (113), Palantir (73), OpenAI (39), Scale AI (13) and Cohere (12). 43 of the 185 companies have any; 142 have none.

4. Small companies: no sign of a gap in our sample

AI-titled share by company board size

We measured company size by how many posts a board has open, which is a hiring footprint, not headcount. AI-titled share of postings:

Board sizePostingsCompaniesAI-titled95% range
Small (under 20 open)3614811.6%8.7% to 15.4%
Mid (20 to 99)3,908747.6%6.8% to 8.5%
Large (100+)16,702639.2%8.8% to 9.7%

Small boards are not below large ones; the ranges overlap, so we cannot say they are above. On Hacker News, 38 of 55 posts that state a small or early stage (69.1%) mention AI, against 8 of 14 (57.1%) that state a large or late stage. Fourteen is too few to lean on.

5. By region

AI-titled share by region

AI-titled share of postings: United States 10.3% (1,112 of 10,751), Europe including the UK 9.1% (343 of 3,782), Australia and New Zealand 6.5% (23 of 353). The Australian sample is small (95% range 4.4% to 9.6%); the US range is 9.8% to 10.9%. The US-Europe gap is small; only Australia is clearly lower. Forward-deployed postings: 185 in the US (1.7%), 78 in Europe (2.1%), 9 in Australia and New Zealand (2.5%).

What we pre-registered, and what happened

QuestionResult that would prove us wrongWhat we foundVerdict
Is it big?Forward-deployed under 1% or AI-titled under 5% of postings1.8% and 9.0%Not disproved
Is it growing?Share flat or falling on Hacker News (under 2 points of change)Forward-deployed +3.1 points; AI/ML-engineer roles -1.0; any AI term +4.8Mixed
Do small companies hire fewer AI roles?Small boards within 3 points of, or above, large11.6% vs 9.2%Disproved in this sample
Is the US ahead?EU or Australia at or above the US10.3% vs 9.1% vs 6.5%Not disproved (US-Europe gap is small)

What surprised us, and what contradicts our view

What these numbers cannot tell you

They are one sample of 185 companies we picked, one snapshot for company boards, and a self-selected Hacker News crowd. Title rules miss roles that do AI work under other names (a spot check found about 7 in 10 unflagged software titles with an AI-like word were real AI work, such as model and inference engineer). The forward-deployed count includes some manager and program-manager titles on forward-deployed teams (7 of 38 in a spot check), and AI-titled includes non-engineering roles such as policy and marketing. Board size is not headcount. None of this measures what small businesses need, only what the companies in our sample advertise.

Method and data

The pre-registration (questions, definitions, and what would prove us wrong), the method notes and the computed results are published with this page. We will repeat the count each month with the same script and publish the change.

Pre-registration (PREREG.md) · How we count (README.md) · Results (results.json)

Sources: company job boards on Greenhouse, Lever and Ashby, collected 12 November 2026; Ask HN “Who is hiring?” threads, November 2025 to November 2026; Indeed Hiring Lab; Fortune.