Medical LLM Prototype Multimodal Clinical AI (Pet Project)

Together with two heart surgeons and a few medical officers, built a prototype deploying LLMs at scale (LLM production systems) that integrates diverse medical data sources such as MRI, CT, X-ray, ECG, and EEG signals into a unified platform. This initiative was piloted with Tampa General Hospital and Mayo Clinic to demonstrate how advanced AI could support clinical workflow automation patterns.

The app highlighted the ability to process historic patient data, align it with current findings, and generate actionable insights. By combining imaging, cardiac signals, and structured medical records, the system showcased how multimodal data could be harnessed to improve patient care and streamline diagnostics.

During the demonstration, the app simulated real-world scenarios where nurses and doctors accessed a patient's historical data to guide treatment decisions. For instance, a nurse could quickly review prior ECG patterns to monitor changes in cardiac health, while a physician could compare MRI or CT results over time to track disease progression. The LLM production systems served as a conversational interface, enabling clinicians to query the system naturally and receive context-aware responses underscoring the potential to reduce cognitive load, enhance efficiency, and deliver more personalized and timely medical services.