This blog post originally appeared on Luca Fury's website.

Now is one of the best times to be a Ph.D. level computer scientist with the skills needed to build artificial intelligence. Only about 20,000 of them exist in the world, and the demand far exceeds that supply. To recruit these uniquely qualified candidates, companies are paying stratospheric salaries, averaging upwards of $300,000 per year.

For startup ventures, a market that competitive presents a number of challenges. With limited capital, it is difficult for them to compete. Such high salaries are also a bit too much to risk on a single candidate. What if the new hire leaves after a short time? The new hire could also lack the skills and abilities the company thought it could count on to drive their AI program forward.

Startups just have less capacity to absorb big salaries but, without the talent, they get left behind. Tech entrepreneurs are not giving up on recruiting the best and the brightest, however. They are resorting to creative and aggressive tactics to get the talent on their teams.

AI needs scientists who are able to build platforms for facial recognition software, self-driving cars, virtual assistants, and other cutting-edge developments. Increasingly, startups are finding that they need so-called unicorns, which are candidates with extremely unique abilities, experience, and personalities.

In recent piece featured in Inc. Magazine, the case of deep learning startup CrowdAI was highlighted, and Founder Devaki Raj explained how startups differ greatly from the environments most AI experts come from. Many candidates have academic backgrounds. They are used to big institutions and often feel more comfortable with the big company structure. Startups, Raj notes, are more ad-hoc. A successful candidate needs the ability to not just develop AI but also develop it in the startup atmosphere.

Another example of such strategies can be found with New York-based 

FuseMachines. Founder Sameer Maskey explained how he fought bigger companies for AI talent and often lost out simply because he lacked the resources to compete. Maskey got creative. He spent 4 1/2 years building a team of over 100 AI engineers who work remotely from around the globe, including places as close as Canada and as far flung as Nepal.

Raj, too, has gotten creative, bringing flowers and balloons to a candidate who was in the hospital. That sent the message that working for CrowdAI would be a far more personal experience than one of the big tech firms. In a market as competitive as AI, sometimes a little creativity gets employers further than a salary bump.