Most of the hiring news I've read this month has been gloomy. Here's a quieter data point that cuts against the gloom: demand for AI engineers is still active, and it isn't coming only from the companies you'd expect.
One caution up front. This is a snapshot from a single job aggregator, not a labor-market statistic. Weekday's U.S. listing showed 519 open AI and AI engineer roles as of October 14, with 166 added in the previous seven days. That means roughly a third of the open roles were posted within the last week. Postings are not hires, and aggregators count differently than the BLS does. Read it as a signal of direction, not size.
Who is posting
The employers with the most open roles on that page: UnitedHealth Group with 17, Waymo with 9, Atlassian with 7, and CLERAD and Coupang with 5 each.
A health insurer is out-posting Atlassian. A robotaxi company also out-posts Atlassian. If you've been assuming your AI hiring competition is other software companies, that assumption is out of date. Healthcare, mobility and logistics are pulling from the same pool of machine learning and AI engineers you are, and some of them have budgets and mission pull that a mid-sized software shop can't match.
The pay spread is wide
The page shows 20 listings at a time, and six of the 20 I looked at disclosed no salary. The ones that did ran from about $71,000 to above $300,000. Some examples:
- Principal Machine Learning Engineer at ZoomInfo: $193K to $303K
- Director of AI and Machine Learning at Thorn: $188K to $302K
- Machine Learning Engineer, Safety, at fal: $180K to $250K
- Senior AI Engineer at QXO: $150K to $220K
- MLOps Engineer at World Wide Technology: $104K to $130K
Several postings, including a Senior AI/ML Engineer role at UnitedHealth Group, listed no salary at all.
Here's what I think: "AI engineer" isn't one job anymore. The title now covers a principal-level engineer and a pipeline operator, and the market prices them very differently. If your req says "AI engineer" and your budget is a single number, you're probably describing two different people and will lose both searches.
What I'd do about it
Split the role before you post it. Do you need someone to design and train models, someone to integrate and ship them, or someone to keep them running in production? Those are different candidates, with different pay bands and different places to find them. The World Wide Technology MLOps range isn't a bargain version of the ZoomInfo principal range. It's a different job.
Post the pay. About 3 in 10 of the listings I looked at didn't, UnitedHealth Group's among them. Picture the candidate with two tabs open: one shows a range that tops $250K, the other shows a blank. The blank loses. A disclosed range also keeps you honest internally about whether the role is fundable.
Stop benchmarking only against software companies. Ask your recruiter which non-tech employers are hiring for your profile (the answer may surprise you). The engineer who left a SaaS company last year may be sitting at an insurer like UnitedHealth Group or a mobility company like Waymo now.
Use contract talent for the operational layer. Pipeline work and data plumbing are often scoped well enough for a contract engagement. Deployment and monitoring often are too. Keep the folks who set your model strategy and own your data on payroll.
And move faster than your process. With a third of postings added in the past week, good candidates have options fast. A six-week interview loop for a scarce candidate is a decision to lose them.
The takeaway
Most of this month's hiring news has been gloomy. In one corner of the market, it clearly isn't. If you need AI engineers, your competition is broader than you think and the pay range is wider than your req admits. The window is short.
If you're trying to fill an AI, ML or MLOps role and aren't sure how to scope it, VC5 Consulting can help you sort out what to hire permanently and what to staff on contract.
Source: Weekday, "AI & AI Engineer Jobs in USA", accessed October 14, 2026. Figures are a job-board snapshot and will change.