Alongside our products, we explore how AI, healthcare data infrastructure, and reliable software can make healthcare systems more capable, connected, and dependable.
Our interests include clinical AI agents, healthcare interoperability, reactive FHIR infrastructure, and methods for evaluating intelligent systems in real healthcare environments.
Healthcare systems are complex, fragmented, and constantly changing.
A clinical task might depend on new observations, changing availability, updated medical records, and operational rules.
An AI agent may produce a reasonable plan. That doesn't mean the plan stays valid when the underlying information changes.
Healthcare AI needs robust infrastructure, meaningful evaluation, and careful integration with clinical and operational systems.
How can AI agents interact with healthcare systems, manage multi-step workflows, and operate within clinical and operational constraints? We explore agent architectures, tool use, task execution, human oversight, and the practical boundaries of automation in healthcare.
How do we know whether an AI system is reliable enough to perform consequential healthcare tasks? We're particularly interested in evaluating agents under changing clinical and operational conditions, including whether actions remain valid against authoritative system state when executed.
How can healthcare information systems better support modern applications and AI? Our interests include FHIR, interoperability, real-time data access, reactive application architectures, and infrastructure designed for intelligent healthcare systems.
AI can interpret information, communicate, reason about workflows, and handle operational variation.
But healthcare also requires explicit rules, permissions, validation, consistent state, auditability, and human accountability. Our approach is to combine these capabilities rather than assume AI should independently control every part of a healthcare workflow.
Open standards, shared tools, reproducible evaluations, and collaborative research are essential to building better healthcare technology.
Members of our team also contribute to healthcare AI research and open-source initiatives through HAIHQ, an independent nonprofit organization.
Its work includes healthcare AI tooling, evaluations, and infrastructure initiatives such as ReFHIR.
Explore HAIHQWe're interested in collaborations with researchers, healthcare professionals, universities, engineering teams, and organizations studying reliable AI and connected healthcare infrastructure.
Get in touch