AMD Silo AI and Delphyr AI Collaborate to Bring Practical, Scalable AI to Clinical Care

The collaboration focuses on healthcare AI, helping clinicians access fragmented patient information faster and within existing clinical systems

Healthcare generates vast quantities of data, yet clinicians frequently face challenges in accessing this information, resulting in decisions made with incomplete or fragmented data. Delphyr AI addresses this challenge by helping healthcare professionals instantly retrieve, understand, and act on patient data spread across Electronic Health Records (EHRs). At the same time, health care organizations do not need to replace their existing infrastructure to benefit from these capabilities. In collaboration with AMD Silo AI, Delphyr is now scaling its platform for clinical settings, where speed, privacy, and reliability are essential.

Addressing the core requirements for healthcare AI

While large language models have shown strong potential in healthcare, successful deployment depends on more than model quality alone. AI must work within existing EHR systems and workflows while respecting the healthcare organizations’ need for privacy, predictability, speed and control. This collaboration focuses on those requirements at the system level, helping bridge the gap between AI potential and clinical usability.

“Healthcare AI must be both safe and reliable, without compromise. At Delphyr, we’re among the few bringing agentic AI from demos into real clinical production. Our collaboration project with AMD Silo AI shows that it is possible to deliver high performance while maintaining the trust and safety clinicians depend on, directly within their existing workflows”, says Michel Abdel Malek, CEO and Founder, Delphyr AI.

Co-development drives optimal AI for clinical use

Delphyr AI and AMD Silo AI are working together to enable healthcare professionals to use AI in their daily work, so they can spend more time on patient care and less on administration. The collaboration centers on two priorities: building high-performance embedding pipelines and refining workflows based on real clinical use.

Co-development is at the core of this cooperation. More than providing infrastructure, AMD Silo AI's engineers work side by side with Delphyr to co-architect and optimize embedding pipelines on AMD Instinct™ accelerators. This work includes support for workload-specific tuning, ensuring that end-to-end system performance meets real clinical demands. For healthcare organizations this means faster time to production, expert support proven on their real workloads, and a system co-designed for their scale, not generic benchmarks run in isolation. 

Built for seamless integration, Delphyr’s platform fits naturally into existing clinical workflows, allowing healthcare professionals to search and summarize patient data, access relevant guidance, and automate documentation without changing current systems. Its workloads are tuned on AMD compute platforms, using the ROCm™ software stack enabling faster clinical information retrieval. 

About Delphyr AI

Delphyr develops AI assistants for healthcare with the aim of giving healthcare professionals more time for direct patient care by helping them retrieve, understand, and act on patient information EHRs. Its platform is designed to integrate with existing hospital infrastructure, enabling better information access and workflow support without requiring system replacement.