Exclusive Content:

Haiper steps out of stealth mode, secures $13.8 million seed funding for video-generative AI

Haiper Emerges from Stealth Mode with $13.8 Million Seed...

Running Your ML Notebook on Databricks: A Step-by-Step Guide

A Step-by-Step Guide to Hosting Machine Learning Notebooks in...

“Revealing Weak Infosec Practices that Open the Door for Cyber Criminals in Your Organization” • The Register

Warning: Stolen ChatGPT Credentials a Hot Commodity on the...

Scaling Patient Care: How Omada Health Enhanced Services with Customized Llama Models on Amazon SageMaker AI

Transforming Nutrition Guidance: The Power of AI at Omada Health

Co-Written with Sunaina Kavi, AI/ML Product Manager at Omada Health


In this article, we explore the revolutionary launch of OmadaSpark, an AI-driven nutrition solution designed to provide personalized nutrition education and support for members managing chronic conditions. We’ll delve into the technical collaboration between Omada, AWS, and Meta, the innovative features of OmadaSpark, and its significant impact on healthcare delivery.

Revolutionizing Nutrition Education: The Launch of OmadaSpark

This post is co-written with Sunaina Kavi, AI/ML Product Manager at Omada Health.

In 2025, Omada Health, a pioneer in virtual healthcare delivery, took a bold step forward by launching a new nutrition experience called OmadaSpark. This innovative AI-driven tool was designed to empower members by delivering real-time motivational interviewing and personalized nutrition education, all built on AWS infrastructure.

OmadaSpark isn’t just another app; it aims to help users recognize their own motivational challenges—such as emotional eating—make informed food choices, set achievable goals, and sustain meaningful behavioral change. The following screenshot showcases the personalized approach OmadaSpark takes in delivering nutritional guidance, allowing users to interact seamlessly and receive tailored advice.

The Opportunity for AI-Powered Nutrition Guidance

Nutrition education lies at the heart of Omada’s chronic condition management programs. While health coaches provide invaluable personalized support, the increasing demand for instant nutritional information presented an opportunity for technological enhancement. Omada recognized the need for an innovative solution that would supplement their coaches’ expertise, enabling them to focus on deeper, more meaningful member interactions.

OmadaSpark is crafted to identify real-world emotional and practical barriers that hinder healthy eating, particularly in an environment flooded with ultra-processed foods. Utilizing motivational interviewing techniques, OmadaSpark employs thought-provoking questions that encourage members to identify their goals, empower their autonomy, and motivate habit changes. With features like nutritional education, tracking capabilities, barcode scanning, and photo-recognition technology, OmadaSpark offers flexible and non-restrictive support, fostering a healthier relationship with food.

“Our vision is to view AI as a force multiplier for our health coaches, not as a replacement,” explains Terry Miller, Omada’s Vice President of Machine Learning, AI, and Data Strategy. “The collaboration with AWS and Meta enabled us to create an AI solution aligned with our values of personalized, evidence-based care.”

Solution Overview

Omada Health developed the Nutritional Education feature using a fine-tuned Llama 3.1 model on Amazon SageMaker. Through this collaboration, they implemented the model following the Quantized Low-Rank Adaptation (QLoRA) technique, which optimizes learning from smaller datasets. Initial training comprised 1,000 question-answer pairs derived from Omada’s internal care protocols and peer-reviewed literature.

High-Level Architecture:

  1. Data Upload: Nutritional education datasets are uploaded to Amazon S3 for training.
  2. Training: Utilizing Amazon SageMaker Studio, the Llama 3.1 8B model is fine-tuned. Resulting model artifacts are saved to S3.
  3. User Interaction: Through the mobile client, user questions invoke requests for personalized nutrition information, taking into account member profiles and conversation history.
  4. Generative Response: Personalized nutrition education is then relayed back to the mobile client, enhancing the member experience with evidence-based advice.

To ensure optimal model performance, Omada uses LangSmith, an observability service that captures inference quality and conversation analytics for ongoing improvements. Registered Dietitians then conduct human reviews to verify clinical accuracy, ensuring members receive safe and accurate information.

Collaboration and Data Fine-Tuning

A key factor in Omada Health’s successful AI implementation is the close collaboration between clinical experts and AI developers. Sunaina Kavi highlights the importance of this partnership, stating:

“Our collaboration with the clinical team was pivotal. It helped establish trust and ensured the model was optimized for real-world healthcare needs. By working together on data selection and evaluation, we ensured that OmadaSpark not only provided accurate information but also upheld the highest patient care standards.”

Patient data protection was of utmost priority throughout the development. Model training and inference occurred in a HIPAA-compliant AWS environment, ensuring that sensitive data remained secure while retaining the necessary control for healthcare applications.

Omada’s rigorous testing protocols involved regular human reviews of model outputs, launching the entire workflow in just 4.5 months. Continuous monitoring and iterative fine-tuning based on real-world feedback ensured sustained performance and member satisfaction.

Business Impact

The introduction of OmadaSpark significantly enhanced member engagement. Users of this AI tool were three times more likely to return to the Omada app than those who didn’t engage with it. This innovation dramatically reduced the time it took to respond to nutrition inquiries, shifting from days to mere seconds.

Building on this success, Omada is deepening its partnership with AWS and Meta to expand its AI capabilities, including fine-tuning models, optimizing context windows, and broadening its scope to additional health domains.

“Our partnership has highlighted the immense value of strategic collaborations in healthcare innovation,” says Miller. “As we look ahead, we’re excited to evolve our platform to offer even more robust support for our members.”

Conclusion

Omada Health’s implementation of OmadaSpark exemplifies how healthcare organizations can effectively integrate AI while addressing specific industry requirements and member needs. By leveraging Llama models on SageMaker AI, Omada enhances the crucial human element of health coaching and enriches the member experience.

The collaboration between Omada, AWS, and Meta stands as a model for how organizations in regulated industries can swiftly and successfully develop AI applications using innovative models on a secure, trusted infrastructure. This endeavor not only transforms care delivery but also reinforces the personalized, human-centered approach that makes Omada effective.

“This project proves that responsible AI adoption in healthcare is not just possible—it’s essential for reaching more patients with high-quality care,” concludes Miller.

As Omada continues to innovate, their commitment to integrating human care with AI-driven technology remains steadfast, paving the way for enhanced support, confidence, and autonomy among members.


About the Authors

Sunaina Kavi is an AI/ML product manager at Omada, focusing on leveraging AI for behavior change in managing diabetes, hypertension, and weight loss. With degrees in Biomedical Engineering and an MBA, she combines technical and business expertise to drive impactful innovations.

Breanne Warner is an Enterprise Solutions Architect at AWS, dedicated to enabling healthcare and life science customers to harness generative AI efficiently on AWS.

Baladithya Balamurugan works as a Solutions Architect at AWS, specializing in ML deployments and helping customers optimize their ML solutions via Amazon SageMaker.

Amin Dashti, PhD, is a Senior Data Scientist at AWS, applying his extensive research background to enhance model customization and training.

Marco Punio is a Senior Specialist Solutions Architect focused on GPU-accelerated AI workloads on AWS, with a passion for high-performance computing.

Evan Grenda works at AWS, collaborating with third-party foundation model providers to solve enterprise agentic AI challenges effectively.

Latest

JD Sports to Enable Purchases via AI Platforms | JD Sports Fashion

JD Sports to Enable One-Click AI Purchases Through ChatGPT...

Is Kraken Robotics (TSXV:PNG) Overvalued Following Its Year-Long Surge?

Is Kraken Robotics at C$8.43 Overvalued After Its Recent...

RegTech Landscape: Navigating Regulatory Challenges, Automation Trends, and Market Potential

Transforming the Landscape of Compliance: The Future of the...

Sir Henri Deterding’s Unexpected Return: Powered by Generative AI – Royal Dutch Shell Plc

Captivating Dialogue: A Darkly Humorous Exchange Between Corporate Titans By...

Don't miss

Haiper steps out of stealth mode, secures $13.8 million seed funding for video-generative AI

Haiper Emerges from Stealth Mode with $13.8 Million Seed...

Running Your ML Notebook on Databricks: A Step-by-Step Guide

A Step-by-Step Guide to Hosting Machine Learning Notebooks in...

VOXI UK Launches First AI Chatbot to Support Customers

VOXI Launches AI Chatbot to Revolutionize Customer Services in...

Investing in digital infrastructure key to realizing generative AI’s potential for driving economic growth | articles

Challenges Hindering the Widescale Deployment of Generative AI: Legal,...

Designing TrueLook’s AI-Driven Construction Safety System Using Amazon SageMaker AI

Elevating Construction Safety with AI: A TrueLook and AWS Collaboration Unleashing Real-time Jobsite Intelligence Through Cutting-Edge Technology The Critical Challenge of Construction Safety Innovative Solutions for Automated...

How Beekeeper Enhanced User Personalization Using Amazon Bedrock

Navigating the Evolution of Large Language Models: Beekeeper's Dynamic Solution for Frontline Workforce Optimization Co-authored by Mike Koźmiński from Beekeeper In this article, we explore...

Cross-Modal Search Using Amazon Nova Multimodal Embeddings

Unlocking the Power of Crossmodal Search with Amazon Nova Multimodal Embeddings Bridging the Gap between Text, Images, and More Exploring the Challenges of Traditional Search Approaches Harnessing...