At January AI, I architected the deep learning framework and data pipelines behind blood glucose prediction, then developed its production deployment for real-time inference. I also engineered a zero-shot system that predicts glucose responses without hardware-derived baseline data, recognized with the CES Innovation Award 2025.
I built a scalable RAG pipeline for Mirror health AI, including data ingestion and serving across more than 50,000 clinical articles. For glucose forecasting, I addressed real-world gaps in user scanning with simulated delays and designed targeted data masking to support generalization.
As Lead Data Scientist and lead author on research published in Nature Digital Medicine, I worked on data analysis, statistical validation, and the foundational meta-learning framework. My work also includes benchmarking food-scanning technology against LLMs and developing a virtual continuous glucose monitor.
