Johnson & Johnson bets on data science powered by Nvidia

An article on the Nvidia blog shows pharmaceutical giant Johnson & Johnson injecting Data Science into its business to improve manufacturing, clinical trial enrollment, prediction, and more. “I really like to call it decision science,” Jim Swanson, corporate executive vice president and corporate information manager, said in a panel discussion at the latest Nvidia GPU Technology Conference. “It’s not just about creating a model – it’s actually the decisions and ideas that you try to extract from those models.”

Machine learning processes, known as MLOps, are a set of best practices for companies to successfully perform AI. Through the MLOps strategy, companies can leverage data to answer challenging questions and drive business operations in a measurable way.

In addition to integrating more than 1,000 data scientists into its business, Johnson & Johnson is building digital literacy across the enterprise—helping employees understand the potential of machine learning models at work.

For example, Johnson & Johnson has formed a Data Science Council internally and developed what Swanson calls “bilingual data scientists” — roles that combine deep domain knowledge with data science skills. “They have this understanding of the main business problem and they have the skills to do data science,” he said.

Swanson said this strategy helps integrate the company’s data science community into business workflows, enabling faster application of machine learning, feedback, and impact models. It also helps overcome reluctance to embrace data science.

“You really have to show them by demonstrating over and over again that this model doesn’t replace the valuable soft skills assets that you have in your field, it offers longitudinal perspectives that you can’t get on your own,” he emphasized.

As companies increase their adoption of MLOps, they need a robust AI infrastructure to support their engineers and data scientists, said spokesperson Nick Elprin, CEO of Domino Data Lab. Engineering is trying to build these tools on its own, which is more difficult,” he said, recommending that companies start with a third-party platform like Nvidia GPU Accelerated Domino Data Lab. “Your engineers will create much more value when they focus on issues that are competitively different and unique,” ​​he said. It has your business.”

To help companies get started with MLOps, Nvidia offers a suite of open source tools on the NGC Software Hub for managing AI infrastructure based on Nvidia DGX systems. Healthcare companies and medical researchers are using DGX systems and the Nvidia Clara application framework to power healthcare models that analyze electronic medical records, increase computational drug discovery, and power AI-enabled medical devices and imaging workflows.

In addition to integrating more than 1,000 data scientists into its business, Johnson & Johnson is building digital literacy across the company—helping employees understand the potential of machine learning models on the job. Swanson gives the example of an indoor hackathon event conducted by Johnson & Johnson to improve insight into the eye care business. By better predicting how many customers will need each product in the Acuvue contact lens line, the company can more efficiently manufacture the products people want. Swanson said that for every percentage point Johnson & Johnson improves forecast accuracy, the company increases revenue, “because you have the right product in the market.”

Dozens of teams across the company — most outside the eye care industry — signed up for the hackathon, which used Domino Data Lab’s MLOps Enterprise platform. The buying team came up with the best model. “We solved a really big problem with real impact and learned some new tools that they didn’t know before,” Swanson said. “With a simple project aligned with a truly meaningful outcome, you can do amazing things,” he commented.

service
www.nvidia.com