Gradient Health

Gradient Health is a medical technology company that connects healthcare data holders with AI developers, providing secure access to de-identified multimodal medical data, including medical imaging. The company sources data from hospitals, clinics and health systems, de-identifies it to remove protected health information, curates and ingests records into its Atlas platform, and delivers selected datasets to customers for download in as little as 48 hours.

The platform is aimed at training, validating and improving clinical AI models. Features described include machine-learning-ready formatting, search by disease and imaging modality, security controls for compliance with patient-privacy requirements, and an AI Search Assistant that translates natural-language dataset requests into structured search filters. The company also maintains public engineering repositories covering DICOM tooling, including a library for storing and interacting with DICOM files in the cloud and forks of the OHIF Viewer, Cornerstone3D and pydicom.

Gradient Health was founded in 2018 and has been described as headquartered in Raleigh, North Carolina and in Durham, North Carolina. Leadership includes Joshua Miller (Chief Executive Officer), Benji Meltzer (Chief Data Officer), Ouwen Huang (Chief Scientific Officer), Pia Spratley (Chief Product Officer), Rod Greco (Chief Revenue Officer) and Katelin Buckmaster (Chief Compliance Officer and Lead Counsel), with a board including Jay Goss, Matt Bailey, Satish Tadikonda and Shana Barghouti.

Founding story

Gradient Health was founded in 2018 to address the difficulty healthcare AI teams face in sourcing diverse, representative, high-quality medical data, bridging the gap between AI developers and healthcare organizations that hold clinical data.

Business model

B2B data-as-a-service: Gradient Health aggregates and de-identifies medical data obtained from healthcare provider partners and makes curated datasets available to AI developers through its self-service Atlas platform. Caplight characterizes the customer profile as data-as-a-service with revenue share, indicating that participating data partners share in the economics of data made available through the platform.

Revenue is generated from providing AI developers with access to curated medical datasets, with a data-as-a-service and revenue-share structure involving the healthcare organizations that contribute data.

Traction

The company states it has millions of patient journeys available on its platform and cites customer testimonials from Riverain, Mango Medical and RapidAI. External firmographic data lists roughly 46 employees, and website traffic signals reported 4.1K monthly visitors in June 2025 and 6.4K in March 2026.

Latest developments

Recent company updates include the launch of an AI Search Assistant in the Atlas platform and published material on data quality in medical AI development and the role of linked, representative clinical data in multimodal foundation models.