Health AI doesn’t fail because of a lack of ideas. It fails when teams hit the reality wall: real-world data, real-world constraints, and not enough compute to iterate fast, validate properly, and ship safely.
That’s why AI4Health.Cro was proud to be featured at the 3rd Day of the Croatian Competence Centre for High Performance Computing (HR HPC CC), hosted by SRCE (University of Zagreb Computing Centre) in Zagreb on 11–12 November 2025.
Spotlight: “When the public helps doctors: the AI4Health.Cro challenge”
In the programme block dedicated to real HPC use cases, Dr. Anja Barešić, project coordinator from the Ruđer Bošković Institute, delivered a talk titled “When the public helps doctors: the AI4Health.Cro challenge”, showing how high-performance computing underpins practical innovation in healthcare—and sharing lessons learned from an AI4Health.Cro hackathon format.
The presentation highlighted how crowdsourcing and challenge-driven development can help close a persistent gap in healthtech: the need for up-to-date AI models trained and evaluated on real data from the Croatian healthcare system.
A key example: the challenge workflow for breast lesion detection in mammography images, including the technical setup, dataset handling, and why compute capacity is essential for running serious experimentation cycles—not just one-off proofs of concept.
Why HPC matters for healthtech teams, and not only for researchers
For startups and product teams building in medical AI, HPC is not “nice to have.” It’s what makes the difference between:
-
slow iteration vs. rapid model testing and benchmarking
-
fragile experiments vs. reproducible pipelines
-
a promising prototype vs. an implementation-ready tool
The HR HPC CC event itself reflected that direction: it combined a dedicated “HPC for entrepreneurs” workshop (focused on practical access to HPC resources and hands-on work on advanced compute infrastructure) with a programme designed around Europe-wide HPC initiatives, real use cases, and a panel on strengthening digital and AI capability.
AI4Health.Cro: building with the infrastructure reality in mind
AI4Health.Cro’s approach is simple: start from validated healthcare needs, build with implementation constraints in mind, and connect innovation teams with the ecosystem required to scale, clinical partners, data readiness, evaluation pathways, and the compute that makes serious development possible.
If you’re building at the intersection of medicine and AI, this is the standard we’re betting on: less hype, more deployable capability—with the technical backbone to prove it.