Advancing Healthcare With Dataiku and NVIDIA

Scaling AI Brad Genereaux, Kelci Miclaus, and Shashank Gaur

Last month, in an insight-packed webinar, Kelci Miclaus, Global Head of Healthcare and Life Sciences AI Solutions at Dataiku, Shashank Gaur, Solution Engineer at Dataiku, and Brad Genereaux, Global Lead Healthcare Alliances at NVIDIA, shared and discussed the state of AI in healthcare and how NVIDIA AI services are helping organizations in unlocking digital health and medical imaging workflows, followed by demonstrations of seamlessly using NVIDIA AI services in Dataiku. 

Go ahead and watch the session, or keep reading for the top highlights and takeaways. 

→ Watch the Full Session Recording

Introduction

The healthcare industry is on the cusp of a technological revolution, with AI poised to transform patient care and outcomes. In this joint webinar, experts from Dataiku and NVIDIA showcased how their partnership is enabling organizations to unlock the full potential of AI in healthcare and life sciences, particularly with digital health and medical imaging workflows.

State of AI in Healthcare

The amount of healthcare data is growing at an unprecedented rate, fueled by the digitization of medical records, imaging, and genomic data. According to RBC, about 30% of the world’s data volume is generated by the healthcare industry. The compound annual growth rate of data for healthcare is estimated to reach 36% by 2025. This data deluge presents both challenges and opportunities for leveraging AI to drive better clinical decision-making. According to a Dataiku survey, healthcare organizations are already adopting AI across the value chain, with notable use cases in research and development, operations, and patient engagement/services.

Kelci highlighted how AI and digital transformations are beginning to enable true prescriptive analytics — understanding the underlying causalities and logic to recommend personalized treatment paths. Together, AI and Generative AI models also have the potential for positive disruption in everything from drug discovery to patient engagement.

state of AI in healthcare

NVIDIA's Services for Digital Health and Medical Imaging

NVIDIA has been a pioneer in the healthcare domain for over two decades, powering medical technologies across the spectrum with accelerated computing and AI. The company offers a suite of frameworks and tools to accelerate AI workflows in drug discovery, medical imaging, genomics, and digital health. These include:

  1. NVIDIA MONAI(Medical Open Network for AI): A PyTorch-based framework for building and deploying medical imaging models for CT, MRI, ultrasound, whole slide pathology, digital surgery, and more.
  2. NVIDIA Inference Microservices (NIM): Cloud-ready inference services that simplify the integration of Generative AI solutions at scale, including drug discovery, imaging, and genomics applications.

During the webinar, Brad demonstrated the power of NIM containers for tasks such as generating drug molecules, text summarization, and medical image segmentation using pre-trained models like Llama 3 and Vista-3D.

AI software platform NVIDIA

Dataiku and NVIDIA Integration

Talk Topic #3 Covered by Shashank Gaur

Dataiku's analytics workbench platform is designed to enable data scientists, analysts, and business users to collaboratively build, deploy, and manage data-driven projects efficiently. Dataiku integrates with NVIDIA's hardware and software offerings, enabling users to leverage accelerated compute resources and state-of-the-art AI models within a unified, governed environment.

Key features of the Dataiku-NVIDIA integration include:

  1. Access to NVIDIA GPUs and CUDA environments through Kubernetes clusters (on-premises or cloud-based).
  2. Integration with NVIDIA software offerings like MONAI, RAPIDS, and NIM LLMs within Dataiku's visual and coding interfaces.
  3. Ability to train and fine-tune models using NVIDIA frameworks, with full experiment tracking and model management capabilities.
  4. Deployment of trained models as APIs or batch processing pipelines for production use cases.

dataiku and nvidia architecture

Still looking for more? Be sure to check out our recent blog “I Have NVIDIA, Why Do I Need Dataiku?” for insights on how the two work together to democratize AI and accelerate processing, enable teams to conduct more experiments, use advanced algorithms, and more. 

Hands-on Demonstrations

Shashank walked through two hands-on demonstrations showcasing the power of the Dataiku-NVIDIA integration:

  1. Fine-tuning a MONAI 2D image classification model on NIH chest X-ray data, with visual data exploration, preprocessing, and model training/evaluation steps leveraging NVIDIA GPUs.

2. Building a healthcare market intelligence chatbot using Dataiku Answers with NVIDIA NIMs Llama 3 and Embed QA models, integrated with Dataiku's LLM Mesh offering for curating knowledge bases from healthcare, life sciences, and biotech news articles.

Putting It All Together

The webinar highlighted the synergies between Dataiku and NVIDIA in advancing healthcare AI applications. While NVIDIA provides cutting-edge hardware and software infrastructure, Dataiku's platform enables organizations to orchestrate and govern their AI workflows, from data ingestion to model deployment, in a seamless and scalable manner.

With this partnership, healthcare organizations can leverage the latest AI technologies to drive innovation in drug discovery, medical imaging, and patient care while ensuring adherence to industry regulations and best practices. As the healthcare data deluge continues, the Dataiku-NVIDIA collaboration promises to be a game-changer in unlocking the full potential of AI for improving human health.

You May Also Like

10 Key Insights Every Executive Should Know About GenAI

Read More

4 Strategies That Set AI Pioneers Apart

Read More

How Corios and Dataiku Help Enterprises Migrate From SAS

Read More

Execs Back to School With the Dataiku GenAI Field Trip

Read More