Building AI Takes More Than AI Skills

What do employers look for when hiring for people building with AI?

Workers across the labour market need AI skills, but these skills vary widely by career area, and the skills needed to build AI are different from those needed to use it.

This article is part 1 of a Q&A examining how skills differ between roles that use AI and those that do not. Here the focus is on workers building with AI and, as you’ll see, it takes more than just AI skills.

Which roles are companies hiring for when looking for AI builders?

First, a step back: who is actually building with AI?

Using Lightcast Occupation Taxonomy, I’ve identified occupations focused on building AI, where AI concentration is close to 100%.

The chart below shows how demand for these roles has changed over time in the US labour market. General AI Engineers are by far the most common, followed by Generative AI Engineers, who overtook Machine Learning Engineers in second place in 2026.

AI builder roles are overwhelmingly engineering roles, and demand has evolved alongside new AI technologies.

What skills do they need?

The next chart shows the 20 skills most often mentioned in job postings for the AI-building roles identified above.

Four of these skills are broad AI-related skills, but the rest show that these roles require much more than AI. IT skills such as programming (including Python), computer science and data science, as well as human skills such as communication and leadership all appear in the top 20.

This is a reminder that AI jobs require more than AI skills, and even the people building with AI - those deepest in the weeds of it - need broader domain expertise and strong foundational skills.

What are the fastest growing and declining AI skills needed in these roles?

The final chart shows how demand for specific AI skills has changed over time. The skills needed to build AI are evolving with the technology: the fastest-growing include generative AI, AI agents and ChatGPT, while demand for machine learning and predictive modelling has declined since 2022.

This highlights how AI skills are not fixed and underscore the importance of continuous learning to stay current with the technology.

We often talk about AI as a single technology, but the reality is far more complex. Building with AI requires more than technical knowledge: specialised expertise and human skills are just as important, and arguably more so, as specific AI skills continue to evolve alongside new technologies.

On top of that, AI builders have a different skills profile from AI users. That’s the topic of the next Q&A.