• Developed computer vision algorithms for perception in an autonomous driving product, combining techniques in existing literature with physics-based first principles derivations to create Ghost’s unique approach to self-driving

• Architected the initial prototypes the self-driving system, deriving the underlying math, and materializing its development plan, across the areas of perception, driving decisions and controllers

• Built machine learning pipelines to select data, train and evaluate model performance

• Provided insights to the machine learning systems team to build out Ghost’s custom data infrastructure

• Developed initial HW specs for camera based on physics-based models to provide sufficiently good data for models

• Developed algorithms to measure physiological signals in health sensors both from existing hardware and investigative prototypes, leading to the successful delivery of atrial fibrillation detection and sleep tracking on the Apple Watch

• Deep-dived into sensor physics to enable accurate, effective algorithm development

• Designed studies to root cause extremely rare failures

• Conducted early-stage investigations of new health sensors exploring tradeoffs between user comfort, signal integrity and population level variation

• Spearheaded inter-disciplinary research in application of information theory and signal processing to neuroscience

• Investigated the application of information theoretic measures for inferring the existence of signaling structures in the brain; in particular, how directed information can be used to infer the focus of seizures in epilepsy patients