Bio

Dara Rouholiman is a machine learning research engineer at Stanford AIM Lab, where he develops and evaluates predictive and generative models for anesthesia and perioperative medicine. His current research focuses on LLM evaluation in clinical settings, deep learning for time-series forecasting, and ML-driven perioperative risk prediction using electronic health records.

His work on tool-augmented LLMs for clinical calculations was published in Nature's npj Digital Medicine (2025). Previously, he led ML development at COR, an at-home blood-monitoring device startup (3 patents filed), and co-founded Telesphora, whose opioid overdose prediction model was deployed with the Connecticut Department of Public Health. He holds a B.S. in physical chemistry from UC Santa Cruz and serves as Lead Instructor for Stanford SASI's Healthcare Innovation Internship.