AI Expert
An expert in Artificial Intelligence can use Deep Learning / Machine Learning to dramatically increase productivity, or even to make things easy, that before were considered impossible. She can implement new papers and understand them from base principles.
We’re making this analogy that AI is the new electricity. Electricity transformed industries: agriculture, transportation, communication, manufacturing ― Andrew Ng
Skills
Most often, the AI itself is not the only part an AI Expert has to concern herself with other related tasks.
Solid technical skills
Linear algebra, statistics, and critical thinking in general are required. Papers are dense in technical jargon, mathematics, and required previous knowledge. Real-world experience comes in handy too to have an intuition of what might work better. Keeping an eye on competitions, such as those held at Kaggle, are a good way to stay in touch with the pragmatic results in the field.
Data and data flow
Is the training data correct, or did we make some assumptions while collecting? What kind of biases does it encode? These sort of questions might be crucial at the early stage of a project using AI.
Where and in which format the data is hosted, and processed, is also very important. We do not want AI experts to waste too much time on fixing infrastructure. Will custom computer parts be bought, the models trained on the cloud, reuse a pretrained model provided by a third-party? Will frameworks like Tensorflow or Pytorch be used, or a more bespoke one. Will this be run on the user's device, or on the company servers.
Productisation and the business side
Is it cost effective? Does it have acceptable risks in case of failure? While a recommendation system can recommend the wrong movie, or song, some fields are much more risk-averse. Medical diagnosis, self-driving cars, or analyses that influences government policy might have big consequences for individual lives, and should be tested accordingly.
Managing expectations
Some business have come to expect any dream to come out of AI projects. It is important for the AI Expert, then, to manage expectations. While AI might look like magic, it is definitely not.