Skip to main content
AI

As AI emerges, healthcare seeks data engineers

While AI can quickly answer questions, a good old-fashioned human IT pro is still required to connect the right data to analytical models.
article cover

Natalia Gdovskaia/Getty Images

3 min read

Nasim Eftekhari thinks healthcare could use a few more data engineers—the kinds of professionals who know how to work with the predictive models supporting California’s City of Hope medical center, where she oversees applied AI and data science.

With the help of AI, City of Hope patients can be continuously monitored for surgery complications, like sepsis or infection. The analytical models pull from the facility’s electronic-medical records system, known as Epic.

“These models are continuously monitoring patients, and if there is a risk, if there is something that needs to be done, there is a notification that pops up inside Epic, in a physician’s working environment, and they get to take action,” Eftekhari told IT Brew.

While some AI applications are already in effect at City of Hope, other “moonshot” ones are in the works, like using models to closely watch cells for early signs of cancer.

Whether moon-bound or Earth-bound AI, Eftekhari told IT Brew about the AI-specific skills that are in demand—and how to acquire them.

The responses below have been edited for length and clarity.

Do any IT-specific responsibilities emerge or change as healthcare environments add AI?

When we start developing a predictive model, we work with our IT partners to find where this data lives…For example, the data that is coming from devices in the ICU. It might be on a server somewhere. How do we reliably get access to that data? We do a lot of coordination in terms of data engineering interfaces. For a lot of our models, we need to read from Epic in real-time.

As healthcare environments become more AI-equipped, what IT-specific skills are most in-demand

Top insights for IT pros

From cybersecurity and big data to cloud computing, IT Brew covers the latest trends shaping business tech in our 4x weekly newsletter, virtual events with industry experts, and digital guides.

I personally believe you have a lack of data engineers in healthcare. Having data engineers and machine-learning engineers to take care of that entire pipeline, going from data to information to insight back into physicians’ or care teams’ working environment—I think that is something that is not there, right now. We had to develop this whole pipeline in house ourselves.

Another thing that is very sorely needed is MLOps—in all industries, but especially for us…How do you make sure that these models are being retrained on a regular basis, making use of the most recent data or maybe tossing away data that is not very helpful anymore?

Where do you enhance your skills in these fields?

There’s a lot of good online resources that are available for free, or very little cost, on platforms. On Coursera, for example, you can get a specialization in data engineering. There are many machine learning engineering courses. A lot of these skill sets are in common with software engineering.

If you’re talking to younger people, these courses are now being offered in computer science and maybe computer engineering [programs]. There are data-science-specific masters or PhDs now emerging in universities.

What’s the most exciting possibility to you for using data and AI to support healthcare environments?

My professional dream is to contribute to extending world-class healthcare to less privileged individuals through technology…If AI could collect all of the data based on all of the historical data of the patients in the US and how they were treated here, and could recommend, for example, the treatment plan, I think that would be amazing.—BH

Top insights for IT pros

From cybersecurity and big data to cloud computing, IT Brew covers the latest trends shaping business tech in our 4x weekly newsletter, virtual events with industry experts, and digital guides.