#Socialdistancing makes for a good time to introspect. At this moment, our engineers and analysts are actively constructing a Coronavirus “nest” that sums up all studies related to the virus. While they build up public access to scientific knowledge, we spoke to our co-founder and Chief Executive Officer Kevin Kallmes on the origins of Nested Knowledge and its long-term goals.

Kallmes says that Nested Knowledge started as more as a scientific idea to communicate research and much less as a business enterprise. As an intern at the BRAIN (Brain Research through Advancing Innovative Neurotechnologies®) initiative with the National Institutes of Health (NIH), Kallmes was tasked with summarizing the reported risks and capabilities of invasive neurotechnologies, like brain-computer interfaces and deep brain stimulation. As the data piled up, he realized that simply writing a report could not capture the complexity of the science surrounding these technologies. Since Kevin is a researcher, not a programmer, he created the first nested dataset the only way he knew: he put the most relevant information for every technology in easily reproducible formats, and then hyperlinked them together in nested Excel sheets. Once he finished his project, it then struck him that beyond the NIH, other people were asking similarly complex questions and finding similar barriers to effective communication through written sources. “I was surprised that no one in science had thought of taking the scientific method and applying it to how we communicate everything in science, rather than just to any given experiment,” he says. “It offended me that no one had gathered all of that medical and scientific knowledge in one place, when doing so is really just an exercise in good bookkeeping.”

Over a casual visit to Boston to meet his friend Karl Holub, Kallmes discussed his latest project with his friend asking him if programming could help generalize it. Over the next few hours, both friends mapped a design of what became Nested Knowledge. A few months later, Holub quit his job at TripAdvisor and became the Chief Technical Officer of Nested Knowledge in early 2019. He built the first version of Nested Knowledge, elevating it into a structured and updateable database.

Since then, Nested Knowledge has crossed many milestones. In February 2019, the company released their first network meta-analysis that calculates custom analyses based on live data with every user click. After presenting this concept, the team realized that the “back-end” of the technology was actually even more vital than the presentation. “We realized that we need not just build how science is communicated, but really how it is stored,” Kallmes says. “From last year, we have been on an infrastructure building project much more than a communication company, and we build that infrastructure around client-driven projects.” Each time a client comes up with a question, the team designs the search and gathering process in such a way to answer it again and again in the future with newer, updated data. “Every question builds another piece of the puzzle,” he says. Right now, the team’s biggest projects are testing machine-learning data extraction and tailoring their data-gathering UI/UX to get rid of what Kallmes calls the “activation energy,” the amount of energy required for a new researcher to get used to the interface.

Nested Knowledge has ambitious long-term goals. Kallmes says that Holub’s development team visualizes their success as a literature bar: the rate at which their systems intake publications that add up to the sum of all published clinical knowledge. “We are still in the one percent range. In the next couple of years, I expect us to get four million studies into the database, at which we expect to have the literature on all of the data on any clinical study of interest to anyone,” Kallmes states. “Technologically, we couldn’t do this five years ago. This is late from the perspective of what medicine needs, but it is early from the technological standpoint. That is exciting to us, because the gap between current journal methods and modern analytics means we can increase access to knowledge across medicine with one central effort.”