Google France hosted a hackathon to tackle healthcare's biggest challenges
From improving clinical trials to easing administrative workloads, AI is already changing what's possible in healthcare. To help accelerate this progress, Google France recently brought together 130 experts for a 12-hour hackathon focused on building new medical prototypes using open AI models.
Twenty-six teams used Google's open models — including Gemma, MedGemma and TxGemma — to develop functional prototypes addressing challenges that ranged from improving emergency-room triage to providing better support for oncology patients.
The ingenuity on display reflects a wider movement across Europe to apply AI to medicine and life sciences. To support this movement, Google.org announced a $5 million commitment to organizations using AI to advance European healthcare. This initiative will help local organizations and professionals build stronger digital health ecosystems.
The hackathon explored a wide range of potential solutions. Here are the winning projects that showcase howGoogle's open models can help solve pressing healthcare challenges:
- 1st Place: POIG (Precision Oncology Interface Gemma) is an AI system designed to support the complex decision-making process in oncology. The project demonstrated a scalable solution to a critical need with significant potential for patient impact. Team: Arun Nadarasa, Jonas Gottal, Juraj Vladika, Mohamad Ammar Said, Nathan Brahmbhatt, William Gehin
- 2nd Place: VitalCue transforms smartwatch health data into actionable insights using Gemma. The application helps users identify early signs of health issues, supporting preventative care. Team: Martin Maritsch, Nathan Denier, Nour Ben Rejeb, Patrick Langer
- 3rd Place: AURA is an AI assistant that provides instant, objective triage insights for hospital emergency staff. Built with MedGemma and Vertex AI, it is designed to ease physician burden and reduce wait times. Team: Soufiane Lemqari, Leo Cartel, Laura Sibony, Vyacheslav Efimov
- Honorable Mention: IGT Assist is a voice-controlled solution using MedGemma that allows surgeons to manipulate medical images during procedures. Team: Aymeric Lamboley, Kondracki Maxim, Rahma Ait Ouaret, Safae Hariri
- Honorable Mention: Owma is a platform for biomedical researchers that uses multimodal models to integrate diverse patient data — including cutting-edge spatial transcriptomics — to help accelerate oncology research. Team: Barbara Bodinier, Nathan Bigaud, Pierre-Antoine Bannier, Vincent Cabeli
These projects offer a glimpse into how open models and collaborative innovation can create tangible improvements in healthcare. By supporting the physicians, developers, researchers and organizations behind these ideas, we can help turn powerful concepts into real-world solutions.
