continuation of work at Fitbit, following its acquisition by Google

Signal analysis for fitness wearables, relating to optical heart rate measurement, electrodermal activity measurement, and accelerometer motion analysis. Evaluated alternative signal processing and optical designs. Significantly increased speed of signal processing firmware. Coordinated Research activities relating to two products, including the Fitbit Ace children’s edition.

Image and signal processing algorithm design for Optical Coherence Tomography (OCT) systems used in ophthalmology:Algorithm Development- developed new visualizations and segmentations of 3-D data used by academic clinical researchers and practicing ophthalmologists and optometrists- developed robust fitting of surfaces to the eye for visualizations of pathology (US Patent 7,668,342 and continuations)- devised new ways to visualize focal pathology (US Patent 8,045,176)- devised robust method for finding fovea location (US Patent 8,079,711)- researched new statistical diagnostics for Glaucoma (US Patent 7,798,647)- applied pattern recognition to characterization of pathology- developed new image analysis for prediction of disease progression- helped define appropriate data acquisition and processing for new instrument- evaluated, improved, and integrated externally developed algorithms: benchmarking, refinements, and testingClinical Collaboration- worked closely with Clinical department and university researchers to define study protocols and evaluate resultsExternal Research Liaison- guided and evaluated external algorithm development- developed analysis tools for university clinical researchersTechnical Leadership- reviewed algorithms with colleagues and suggested new approaches- reviewed internal and external Intellectual Property

Signal processing algorithm development for pulse oximetry and other medical technologies: - designed feedforward architecture for adapting filter parameters (US Patent 7,016,715) - developed new Independent Component Analysis (ICA) techniques for motion artifact cancellation (US Patents 6701170, 7079880, & 7343187) - applied pattern recognition to the classification of motion artifacts - contributed to new fetal oximetry algorithm - developed algorithms for detection of extravasation using electrical impedance measurements - developed algorithms for estimation of hemoglobin from IR absorption data - researched unconventional signal-processing approaches to reducing system power consumption

Developed software for interfacing and signal processing with custom multiprocessor system for medical ultrasonic imaging. Real-time implementation of adaptive lattice filter for flow estimation. Designed and presented courses on classical and neural pattern recognition techniques.

Assisted in development of high-level architecture and user-interface for liquid chromatography integrator. Contributed to new integration algorithm.

Tested software for medical linear accelerators and fluoroscopy systems. Coordinated Product Assurance activities on 600C Clinac.