I have have spent nearly 25 years searching for emerging technology talent. I recall looking for Java talent when it had only been in beta for 6 months! It was a bit of a challenge, but finding a solution that matched the candidate market availability was key. 

The AI market is not too different to all those previous 'new technologies'. I've been searching for AI talent for about 18 months and spent some of that time in Toronto, which is a big hub in the AI ecosystem.

So what I have I learnt. Firstly there are a lot of people that have Degrees, Masters and Doctorates in an AI related subject. There are a lot of people either looking for their first role or looking to make a move. The biggest challenge is that the candidate pool doesn't always match the clients requirements.

For some companies, research histories, publications at 'top conferences', a high h-index, research areas and the ability to code at production level are essential. Most candidates are looking for a good team (experience, ability to learn from, values, culture), compute power and data (lots of high quality data!). Some looked at the mix of research and development, ability to eventually put their work in a finished system or product, and the research direction.

What I learnt overall was that clients that took the time to understand what was available in the candidate market, looked at the market intelligence and then agreed the possible solutions, were at least half way there. The clients that professionally and passionately presented their opportunity, put forward their employee proposition, and really listened to the candidates also secured some great talent.

It won't be too long until today's 'unicorns' in the candidate market are available, but as always working with the search company and making decisions based on market intelligence will be the quickest way to hire the best AI talent available in the market at that time.