• Engineered a cross-platform insulin-pump-to-cloud communication debugger using Dart & Flutter to intercept, decode, send, & record over 80 unique message types in real-time within Medtronic’s message transport system.• Developed a recursive algorithm to classify messages into 30+ handshake protocols, linking related messages within the same handshake & flagging erroneous sequences to streamline miscommunication debugging.• Created and demonstrated a new feature for delivering real-time Medtronic app notifications to Android Auto CarPlay and desktop by using Dart, Flutter, and Google Cloud APIs to modify Medtronic’s notification system. • Created and demonstrated an IOS Live Activity feature using SwiftUI & WidgetKit, letting users conveniently view the status of their blood glucose sensor warmup period from their lock screen over the span of 2 hours.

• Designing machine learning-based synthetic patient models trained with 260,000 datapoints in Python using Bayesian Belief Networks to replace confidential medical data, aiding diabetes and chronic disease research.• Teaming up with other researchers, clinicians, and engineers at the Georgia Institute of Technology to seamlessly integrate expertise and innovation for accurate synthetic patient representations.• Improved accuracy of synthetic data by ~15% (in regards to standard deviation of diabetic patient attributes) by repeatedly tuning and testing hyperparameters of Bayesian Network model.• Attended meetings and presented work periodically to Georgia Tech faculty to update on progress.

• Developed official SageVR website in React (1000+ viewers in the first month) to communicate product information to investors & potential customers.• Worked in Unity (C#) to rig 3D character models created in ZBrush and Blender powered by AI chatbot for speech. • Engaged in data annotation for over 300 pages of released public therapist transcripts to train AI model.• Reduced memory consumption by ~15% through code refactoring and memory profiling, resulting in improved application stability and reduced system resource usage in VR.