What 615,000 open jobs on 14,700 company career sites say about hiring in November 2026

Every night a Mac mini asks 14,700 companies for their open jobs. It doesn't scrape HTML. Most companies host their careers page on an applicant tracking system (Greenhouse, Lever, Ashby, SmartRecruiters or Workable), and each of these serves the page from a public JSON endpoint. The crawler calls those endpoints, normalizes the fields and stores one row per open job.

The run on 15 November 2026 returned 615,566 open jobs. These are the numbers I found most interesting, with the caveats that come with them.

1. "Forward Deployed Engineer" has become a standard job title

There are 1,956 open jobs with "Forward Deployed" in the title, at 812 different companies. 552 of them were posted in the last 30 days.

Palantir made the title famous. It has now spread across the AI startup world: 52% of these jobs are on Ashby, the ATS most popular with venture-backed startups, although Ashby holds only 11% of all jobs in the index. The median advertised US pay for the role is $200,000.

2. Junior titles have almost disappeared from software job ads

Among 17,792 software engineering and developer jobs on Greenhouse, Lever and Ashby:

An unmarked title can still be a mid-level or junior role, so this doesn't prove there are no junior jobs. But it does show what employers choose to advertise: one junior-labelled software job for every 24 senior-labelled ones.

3. AI titles are a much bigger share of new software jobs

Of software jobs posted in the last 30 days, 14.7% have AI, ML, LLM or "machine learning" in the title. Of software jobs that have been open for more than 90 days, it's 3.2%.

Caveat: this compares jobs that are open today, not all jobs ever posted. AI roles might also get filled faster, which would remove them from the older group. Still, a 4.6x gap is hard to explain without real growth.

4. Pay transparency: the US versus Europe

Greenhouse returns pay ranges only from a per-job request, and SmartRecruiters and Workable rarely return them, so I measured this on Lever and Ashby jobs only (113,000 jobs). This is the share of postings with a salary range:

In the US, state laws (California, New York, Colorado, Washington and others) require a range in the ad. The EU Pay Transparency Directive had a transposition deadline of June 2026. It requires employers to tell candidates the pay range before the interview, but not necessarily in the ad, and many member states hadn't finished transposing it by the deadline. So far, it hasn't made ranges in job ads common in Europe.

The median advertised US salaries (midpoint of the range, in USD per year) for a few titles:

These samples are skewed toward well-funded tech companies, which are the ones using Lever and Ashby.

5. Old listings that are still "open"

11.9% of open jobs on Greenhouse, Lever and Ashby were first published more than a year ago, and 5.9% more than two years ago. On SmartRecruiters it's 37%, but SmartRecruiters dates are less reliable.

These aren't mostly fake jobs. The biggest group is roles that are always open: one fitness chain has 4,272 "Personal Trainer" postings older than a year, followed by healthcare staffing and retail shift jobs. Only 2.8% are labelled as general applications. If you build alerts or a job board on top of this data, filter by posting date. An "open" job from 2024 is probably not one anyone is actively hiring for.

6. Other numbers

• Remote: 15.1% of all jobs are marked remote. On Ashby it's 52.4%, and on SmartRecruiters it's 3.7%. Which ATS a company uses says a lot about what kind of company it is.
• Concentration: the median company has 9 open jobs. The top 1% of companies account for 38% of all jobs, led by Domino's (24,567 store jobs) and big staffing firms.
• Where: US 226K jobs, India 28K, UK 24K, France 17K, Canada 15K, Germany 14K (counting the two thirds of jobs whose location names a country).
• New this week: 9.7% of open jobs were posted in the last 7 days, and 32.5% in the last 30.

How it's built

• Finding companies: board slugs are collected from public links to boards.greenhouse.io/..., jobs.lever.co/... and similar URLs, and checked weekly.
• Fetching: one request per company per ATS (Greenhouse boards-api.greenhouse.io/v1/boards/{slug}/jobs, Lever api.lever.co/v0/postings/{slug}, Ashby's posting API, and so on). The endpoints are listed in this reference.
• Normalizing: title, department, location, country (ISO code), remote flag, employment type, salary min/max/currency/interval, posting date and URL.
• Safety check: the nightly build refuses to publish if the job count drops by more than 30%. A DNS outage taught me that one.

Data: a free sample of 1,000 jobs from this run (CSV, 612 companies, all five ATSs) is in the repo: samples/job-index-sample-1000.csv.

If you want to query the full index yourself, it's on Apify as Job Search API ($1 per 1,000 jobs) and Greenhouse Jobs Scraper. Hiring Signals returns company-level counts, such as "companies that posted 3+ sales jobs in Germany this month".

What should I measure next month? I'm considering time-to-close by role, and which companies' openings grow or shrink month over month.