Given that the healthcare sector is stovepiped and has strict privacy requirements, an out-of-the-box approach is required for better insights.

Exafluence Inc. addresses this white space issue by building an analytics application on top of MongoDB Inc. and deploying master data management concepts for better decision-making, according to Ravikiran Dharmavaram (pictured), founder and chief executive officer at Exafluence.

“The reason why we used Mongo is that Mongo is the only platform that allows you to build [a] lightweight master data management concept,” Dharmavaram said. “For example, patient data can come from 30 systems, but I need to know which system is the system of record. To do that, I don’t need to invest into large master data management products, I can leverage Mongo. In addition to that, the delivery mechanism of the data, it can be distributed on a mobile, on a cloud, and it can be built as a native app.”

Dharmavaram spoke with theCUBE industry analyst Dave Vellante at the “Supercloud 5: The Battle for AI Supremacy” event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how Exafluence employs data analytics and large language models in the delicate healthcare industry for optimality purposes and how MongoDB fits into the picture.

How data analytics helps generate a longitudinal view of patients

MongoDB enables Exafluence to conduct enhanced data analytics in the healthcare industry because it supports the writing of machine learning programs in an open way, according to Dharmavaram. MongoDB is also JSON-centric and supports time series data.

“We are able to identify the patient profile because of the master data management concepts that can be applied in MongoDB,” he stated. “We are able to create a longitudinal view of a patient, get the patient data and do various levels of analytics on them. We’ll do analytics to the level of doing the risk score assessment, looking at the disease methods, population health. We look at overall what impacts them in their care gap mechanism.”

Through its large language models, Exafluence can conduct protocol mapping and clinical trials easily. As a result, the generative artificial intelligence and LLM approach enables Exafluence to streamline and enhance patient research in the healthcare and pharma industries, according to Dharmavaram.

“We built our own LLM models, and the ability to use Mongo’s vector database to refine our models was wonderful,” he explained. “We built a model for clinical trial protocol. We are also working in the specialty chemical industry to find out new molecular research, to figure out how the molecules are formed with easiest route manufacturing methods.”

Exafluence’s decision to build its LLMs was triggered by the fact that data should be confined to a certain environment in the healthcare sector. For instance, LLMs should not take any public routes to the internet.

“In the space we work in, especially in the pharma industry, research and chemical industry, there are lot of documents that are proprietary and tribal knowledge to the system that they work with,” he stated. “It can’t get out of it … we took a model, open-source model, configured it into their environment and made a proprietary LLM for them. We used Falcon; we used a couple of more Python-based models that we configured.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the “Supercloud 5: The Battle for AI Supremacy” event:

Photo: SiliconANGLE