The Result
Hepha AI co-founder Jin Dongyang took 3rd place (General Track) at the Medprompt Hackathon 2025 — a medical-AI hackathon co-organized by REVOTECH, AI Hong Kong and PolyU Design, with sponsors including Pfizer. The award recognizes innovative prompt engineering for medical AI: making models more practical, effective and accessible in the real world — not just accurate in a research lab.
Why We Compete
Benchmarks and competitions keep us honest. Placing among the top teams is an independent check that our medical-AI capability — the same capability that powers our claims pipeline — holds up outside our own environment, against fresh problems we didn't choose.
Why It Matters for Claims
Claims adjudication lives and dies on medical understanding: reading clinical documents, interpreting diagnoses and procedures, and connecting them correctly. The strength recognized at Medprompt is exactly the strength that lets our solution capture, ground, and adjudicate health and travel claims with evidence a human can audit.
Current, Not Historical
Alongside the AAAI/IAAI-recognized research behind our Medical Knowledge Graph, this 2025 result shows the capability is being sharpened continuously — not resting on a track record.
Applied Under Pressure
Hackathon problems arrive unseen and get solved in days. That is the same muscle a claims pilot demands: adapting proven methods to a new portfolio, fast, with measurable results.
Test Us on Your Claims
The best benchmark is your own portfolio. We'll prove it there.