鶹ý

Why One Hospital Hired an AI Prompt Engineer

— Boston Children's found a "unicorn-type candidate" for its newest clinical tech hire

MedpageToday
A photo of the exterior of Boston Children’s Hospital

In the latest move toward integrating artificial intelligence (AI) into healthcare, Boston Children's Hospital hired one of the industry's first AI prompt engineers last month.

The role likely garnered more attention than any other in the history of the hospital, according to John Brownstein, PhD, the Chief Innovation Officer at Boston Children's Hospital, who led the charge .

"We got many, many applications, some really impressive people, but we were looking for a specific set of skills and background," Brownstein told 鶹ý.

He was focused on finding an individual with "a clinical background who knows how to use these tools" -- someone who had experience coding for large language models and natural language processing, but who could also understand clinical language.

Ultimately, he and his team found Dinesh Rai, MD, who offered the right combination of clinical and technical experience.

"It was not easy to find [someone] -- a bit of a unicorn-type candidate," said Brownstein, who is also a professor of Biomedical Informatics at Harvard Medical School.

Rai became Boston Children's new AI Prompt Engineer at the end of August after a 4-month search. Brownstein said Rai was an "n of 1" candidate "who actually practiced medicine, lived in a clinical environment," and had "successfully launched many applications on top of large language models."

Rai first got into coding after his undergraduate years at Rutgers University in New Jersey, and began working with AI tools in 2014 when he arrived at medical school at the Albert Einstein College of Medicine in New York. He subsequently completed his residency in emergency medicine at Northwell Health in New York City, but the pull toward technology led him to a fellowship in informatics.

After his fellowship, Rai decided to leave clinical practice to work with technology companies as a consultant using natural language processing and generative AI tools, bolstering the skills sought by Boston Children's.

While few people have heard of the term , even fewer know what the job involves. In short, prompt engineering is the practice of creating commands for AI programs with the goal of generating a desired response or output. In its simplest version, a person can ask ChatGPT a question. On the more complex end of the spectrum, a person would use a combination of text commands, coding, and even specified datasets to build AI tools to improve responses.

Rai emphasized that there are varying levels of expertise with prompt engineering: "It's kind of like driving a car, where basically anyone can drive an automatic car, and anyone can go onto ChatGPT ... write some text and get a pretty solid response."

Then, there are "people who know how to drive manual, and there are people who will know different prompting techniques, like chain-of-thought or zero-shot prompting," he continued. "Then you have those F1 drivers who are very intimate with the mechanics of their car, and how to use it most optimally."

Rai said his skillset and experience land him in F1-driver type of AI prompt engineer for Boston Children's. For example, he plans to build new programs using AI tools including ChatGPT. He also plans to figure out the optimal approach to providing data to large language models while automating the evaluation of the responses generated using that data.

Rai is already working with several groups in the hospital to improve their AI capabilities, he said, even though he only started a couple of weeks ago.

"Some of the nuances I bring to the table in terms of being a physician and having worked clinically and understanding really deeply the clinical workflows and how we can implement the technology -- where its limits are, where it can excel, and the quickest way to get things [done]," Rai said. "I'm happy to be able to help with all of that."

Having an AI prompt engineer also will allow Boston Children's to improve communication with its AI partners, and build meaningful relationships with major technology companies, Brownstein said.

In general, Boston Children's has been using AI for years, but the recent growth in available tools and interest have helped bring the need for more AI capability to the forefront, Brownstein added.

"I'm lucky to [be] in an organization that has recognized the importance of AI as part of the future practice of medicine," he said.

He added that as the workforce increasingly begins to use large language model tools in their daily workflows, having a prompt engineer will help ensure "that people are using these appropriately, effectively, responsibly. So we need that expertise."

  • author['full_name']

    Michael DePeau-Wilson is a reporter on 鶹ý’s enterprise & investigative team. He covers psychiatry, long covid, and infectious diseases, among other relevant U.S. clinical news.