How to Interpret What Your AI Engineer is Telling You
You’ve hired your first AI engineer, and communication has been…tricky. It’s not that you can’t get things accomplished or that the work has been shoddy. Instead, it seems like you’re talking around each other in meetings and status updates. The good news is that you can learn what your AI engineer is really saying to you by remembering why you hired that position in the first place. Get your pipeline back on track by understanding what these common sentences mean.
“I don’t have time.”
Data science is an umbrella term that covers a wide area. If you hired your AI engineer and expect them to do all the things, your engineer won’t ever get around to the project at hand. AI engineers aren’t there to troubleshoot your IT department. They don’t need to spend hours working through legacy code.
Your AI engineer is a critical liaison between your data science team and business analysts. They’re responsible for producing your minimum viable product. If your AI engineer doesn’t have the support for gathering and cleaning data, they’ll spend vast chunks of their time just trying to get training started.
All these tasks are related, right? Think about it like this. Your supply chain manager would never get anything done if they were in charge of the finances as well. They liaise with finance to ensure appropriate cost policies, but ultimately, the two positions support each other. Don’t make the mistake of hiring your AI engineer and then sticking them with every job you have in the IT department. Figure out what you can do to take some of these tasks off your AI engineer’s plate.
“My model is training.”
AI is data-hungry. Machine learning gobbles up structured data, but deep learning could take a while because there’s so much that has to process. Once your AI engineer builds the model, it isn’t done. It’s not like flipping a switch and turning the lights on.
Instead, the model has to “learn” in a way adjacent to humans. Once the model is built, it must undergo a series of training sets to learn to process the data in a way that answers your question or performs your required action.
Think about it like this. A musician writes a new piece of music and puts it in front of someone. Just because the music has been written doesn’t mean the person is ready to play it. If that person has never touched a piano, it could take a long time to master the music piece. However, once that person spends time practicing and learning the piano, playing the new piece of music is much easier.
This isn’t saying that it will take years for the model to train, but you must be patient, especially in the beginning. Your AI engineer is responsible for orchestrating the framework and the training, and hopefully has support for other pieces. When the model is training, believe it.
“This project would be better for ML/DL/Something else.”
Everyone wants to use artificial intelligence, but not all projects are created equal. In some cases, the project is far too simple to waste time on deep learning or building a fully fleshed out artificial intelligence program.
Boardrooms want to be in on the AI action. However, it’s essential to take your AI engineers word seriously if the project would be better served with one type of program over another. If you’re analyzing simple, structured data and need black and white results, deep learning is going to be a huge waste of resources. If you need an answer quickly, machine learning could be a better option. If you need to understand precisely why and how your program came to the conclusion that it did, deep learning will be frustrating.
Your projects could be a good fit for deep learning if they look into unstructured data, and you don’t need to know how it reaches conclusions. Take your AI engineer’s lead and help the boardroom understand that just saying “AI” isn’t going to get results.
Learning The Language of Your AI Engineer (and Teaching Yours)
AI is still in its infancy as far as business is concerned, so your AI engineer could still be in research mode. It may take a bit to transition to the needs of the company, so some patience upfront is required. You’ll have to allow your AI engineer to utilize the creativity and drive inherent in the position and make sure there is plenty of support for the mundane details of handling your company’s products and data. If you don’t have a team, look into a SaaS option, so your AI engineer is freer to pursue what you hired them to do.
Read more data science articles on OpenDataScience.com, including tutorials and guides from beginner to advanced levels! Subscribe to our weekly newsletter here and receive the latest news every Thursday.
