Building software at a medical data analytics company is complex enough on a regular day, but doing so during a worldwide pandemic presents novel challenges. We asked Stephen Mead, Chief Engineer at Nested Knowledge to elaborate. Mead handles all of Nested Knowledge’s software, whether it is the website, user interface, functionality, and implementation, from writing the HTML to building the data structures for tables on the database. “I’d say my job is like a full stack developer++, but that’s what happens when you work in a startup,” he says with a smile. Together with Chief Technology Officer Karl Holub and Director of Research Natalie Reierson, Mead works on various features of the software running Nested Knowledge.
One of Mead’s recent projects was redesigning the Nested Knowledge homepage. “I know it looks very simple, like not much is going on. But a lot of it is not visual: a lot of complexity goes on behind the scenes. When you do a literature search, you’re not just getting the study information from PubMed; it’s also running through things like Population and Outcome date extractors for factors like patient subpopulations, intervention, comorbidities, outcomes and things like that.” His other recent projects include gathering data from multiple sources such as the directory of open access journals and redesigning the user tool of the data gathering process. “When you get the core pieces of your product in place, then it’s the little things to refine.”
Interacting with an amazing software team has been one of the best perks of his job, says Mead. He particularly looks up to his colleague Holub. He recalls the first time he saw Holub’s resume and thought: “Wow, yes please. I definitely want to work with this guy.” He says that having committed and accomplished people on the team can charge the atmosphere. “He (Holub) is so smart about the way that he teaches you, and he’s just really good at his job. And everyone can appreciate that. It just makes your life easier and elevates everyone else,” he says.
The flexibility of working from home is also a big plus for Mead. “I love having difficult-to-reach deadlines. I’m up until 4 in the morning working but also knowing that I don’t have to be back at work until 3 in the afternoon,” he says. “Making my own schedule and being responsible for my own deadlines is important to me. It’s not about how many hours you work but it’s all about your output and how you get things done.” He agrees that working from home comes with its own learning curve. “It’s a tough experience to go through, but once we do get through it, we work more comfortably. But it took me some time to realize that I’m not going to sit at a desk every day to work.”
Mead says that Nested Knowledge is poised to evolve more rapidly in the COVID-19 era, with the increasing push to acquire preprint and prepublication data. “The coronavirus has changed the way researchers look at the speed of publication. A year-and-a-half ago, when we told people that we had perpetually up-to-date literature reviews, they said they didn’t care if they were a week out of date. Now, with COVID-19, that’s no longer okay. We need the studies that were published yesterday in our literature review today,” he says.
“The one thing that remains the same is our vision that the medical world needs the data. And 300mb text files are not the data they need. It doesn’t look like anyone has a good solution for this in the long-term, and I think that Nested Knowledge can fix that,” he says. “Sure, the tools that we use will continue to evolve, but that is the core principle, and I want to see us be successful at that.”
Enhancing the user experience is one of the key future challenges for the company, Mead says. “We are making it usable for a wide range of doctors so that they don’t have to spend hours learning new software. We want a person walking in not knowing anything about us and within 45 minutes doing a pretty comprehensive literature review, all within one piece of software,” he says. “We have all the tools to do that, but there is still a learning curve; we want it to be clearer for our users to do it on their own, and we hope to get better at it.”