Which industries are hiring the most artificial intelligence (A.I.) and machine-learning talent? The answer to that question is important to anyone currently learning those skills, as it can determine where to find the best (and most lucrative) jobs.

According to a Dice.com analysis, education and real estate have seen some of the biggest leaps in A.I.-related job postings over the past few years. Since 2021, healthcare and information have also enjoyed a sustained burst of A.I.-related hiring, despite the numbers declining last year. Here’s the chart:

Given all the hype around A.I. and machine learning, it might seem odd to see the job posting percentages dip in critical industries such as finance. Keep in mind, however, that the actual market for A.I. jobs remains relatively small, especially in comparison to ultra-popular job segments such as software development and data science. It doesn’t take the addition (or subtraction) of many job postings on an annual basis to see a broad impact on the percentages.

It’s also important to keep in mind that companies in finance and tech have been hiring A.I. and machine learning specialists for quite some time, while industries such as healthcare are just catching up—thus the mix of notable spikes and declines over the past few years. As A.I. tools become more democratized and easier to use, there’s a greater emphasis on more employees adopting some A.I. skill-sets, which reduces some of the demand for A.I./ML specialists.

Keep in mind that, no matter what the data, there’s a big future for those who know how to build and operate A.I. and machine learning models. For example, O’Reilly estimates the average salary of data and A.I. professionals at $146,000 per year (that’s from 2,778 respondents in the U.S. and 284 in the U.K.); salaries increased an average of 2.25 percent annually. Meanwhile, another recent study of A.I. use-cases showed many companies using artificial intelligence and machine learning in sales, CRM, chatbots, cybersecurity, and marketing automation—pretty much the entire tech stack.