Employers are hiring for AI that survives launch

Tech job postings in the United States were up 20% in October from a year earlier, according to Dice's October 2026 Tech Jobs Report. Postings that call for AI and machine learning skills were up 91%. For a founder, the more useful detail sits one layer down, in what employers are asking those AI hires to do.

What the postings say

Dice, using Lightcast posting data pulled on November 3, 2026, lists skills whose mentions grew more than 200% year over year. Three clusters stand out:

• AI tools and patterns: Claude, GitHub Copilot, agentic AI, AI agents, retrieval augmented generation and responsible AI.
• Code quality: technical debt, maintainability, design reviews, software technical review and defect management.
• Production operations: telemetry, traceability and exception handling, plus AI adoption, enterprise integration and engineering practices.

Dice reads the code quality cluster as employers wanting engineers who can review and maintain AI-assisted code. It reads the last cluster as the work of monitoring systems once they are live and moving AI from pilots into everyday use.

The month-over-month list points the same way. Titles that grew more than 150% from September included Machine Learning Operations Engineers, Machine Learning Data Engineers and Lead, Staff and Distinguished Engineers. Dice also reports healthcare tech postings up 15% month over month, behind software and insurance.

What this does not show

These are job postings, not hires. A requirement in a posting shows what an employer wants to be able to say it is hiring for, not what it ends up doing. Dice's title list is a month-over-month change, and small bases can produce large percentages.

Our interpretation

This section is interpretation, not part of the Dice report. Building an AI demo is now cheap. Keeping an AI feature trustworthy in month six is not. Data shifts, a customer's integration changes, and someone has to explain why an output moved. If postings are starting to ask for maintenance and traceability skills, the market is pricing the thing that was always scarce: people who own a system after launch.

In healthcare technology the cost of getting this wrong is higher, because customers will ask how the feature behaved with their data and who is accountable when it does not.

What this means for hiring

• Ask about the last system the candidate kept running after launch. What did they monitor, what failed, what did they change?
• Separate tool familiarity from ownership. Naming Copilot or a retrieval framework is easy. Explaining a bad output they traced to its cause is not.
• Name an owner before you open the role. An AI feature without a named person responsible for it after launch is a pilot.
• Weigh review skills for senior engineers. Someone who can review AI-written code well is worth more than someone who writes more of it.

If you are planning an AI, data or engineering hire and want to compare notes on what the role should prove, we are glad to talk.

Keith Anderson, Founder

Sources

• Dice, October 2026 Tech Jobs Report, released November 2026, data as of November 3, 2026.