In my role as CTO I set overall technical strategy for VisionGate and balance the concerns between the various technical groups within the company and am the chief architect of the LuCED test for early detection of lung cancer. I have assembled the technical team at VisionGate's headquarters in Phoenix, Arizona which includes a wet-lab for specimen processing and storage, cytology for cell diagnosis, development of cell classification algorithms based on machine learning, pattern recognition and AI, and IT infrastructure for the company. I have also assembled our Cell-CT lab, responsible for collecting data on our devices to support our various clinical initiatives. My principal technical expertise is signal processing, machine learning and data analysis. In VisionGate, I direct development of 3D cell image formation and image processing algorithms and development of automated classifiers for detection of various cell conditions. Working with our company CEO, I help define the overall strategic direction for the company, targeting applications for our device and defining needed resources and setting schedules. A key role also involves interfacing with the FDA around the approval of our clinical applications.

I am a senior technology executive with 20+ years of experience leading AI teams to patented innovations, peer-reviewed publications, and successful, industry-defining outcomes in cancer detection. My training was in electrical engineering. At NeoPath, we built the FocalPoint system, the world's first fully automated system based on AI to detect evidence of cancer in Pap smears. The FDA cleared the system for use with a class III PMA in 1996. Becton-Dickinson purchased NeoPath in 2006. VisionGate created the Cell-CT, the first system to image cells in sub-micron resolution 3D, enabling AI-based cell recognition to match the best in human-rendered cell identification. As CTO, I have translated product requirements into a development program to create the LuCED test to detect early-stage lung cancer and the VesiGuard test to detect bladder cancer - the first commercial applications for the Cell-CT.My record in AI innovation is well established, with 18 AI-related patents and 27 publications.

VisionGate has developed the Cell-CT, which images cells in 3D. 3D imaging comprehensively represents cell features, enabling new, automated, cell diagnostic tests not possible through 2D imaging. I defined the technical strategy and assembled a team of five highly qualified scientists that created 3D cell reconstructions and two new clinical tests.1. Cell 3D Reconstruction and Cell VideosThe Cell-CT produces raw images that are used to create 3D reconstructions. Critical to this effort was the development of techniques to correct for cell motion occurring during data capture. Examples of 3D cell videos may be viewed by searching VisionGate on www.youtube.com.2. Specimen Scoring AlgorithmsThe Cell-CT was created to foster automated recognition of cell type using 3D biomarkers. Automated recognition is accomplished through training of cell scores to match expertly rendered cell diagnoses. This task involved development of special tools to foster cell diagnosis, an IT structure to processcells and algorithms to process the 3D cell reconstruction to segment cell elements of interest and compute features that correlate to the state of the cell. This information is then used to create classifiers whose output closely matched cell diagnoses. - Lung Cancer Detection ­ Our primary emphasis was the development of a test to detect lung cancer cells and precursors in sputum. Our efforts demonstrated sensitivities for the specimen exceeding 90% for nearly 100% specificity. Several patents are now pending as a result of this work. - Lipid Laden Macrophage Index ­ Aspirated fats are taken up by lung macrophages. The degree of uptake expresses the degree of aspiration possibly indicating gastro-esophageal reflux. We developed a test to automatically characterize the lipid content of macrophages in sputum. Accuracy and precision of our measurement was established in a clinical trial showing improved performance when compared to slide-based methods.

I helped develop the FocalPoint system, which automatically screens Pap smear slides for cell abnormalities. Automation of pap-smear slide processing reduces the resources needed for the time consuming and costly process of slide review by a cyto-technologist while preserving the rate of abnormal slide detection. The focus of my career was the development of image processing and classification algorithms to detect slides with abnormal cells. NeoPath, joined with AutoCyte in 2000 and the combined corporation was acquired by Becton-Dickinson in 2006.1. Primary Screener Pap-smear classifiersI managed a group that developed classifiers that determine the if the Pap-smear slide­ should be reviewed by a cyto-technologist or not. Development resulted in an algorithm that provided optimal detection of abnormal slides while maximizing the sort rate. The sort rate is the percentage of slides the system could safely classify as not requiring review by a cyto-technologist. PMA approval was granted by the FDA in May, 1998.2. Technical leadership and management to increase FocalPoint sort rateI managed a group which defined new classifiers to increase the FocalPoint sort rate by 80% while maintaining abnormal detection rates.3. Technical leadership and management to develop Guided Screening for the FocalPointWith Guided Screening, the FocalPoint identifies locations within a Pap Smear that are most likely to contain abnormal cells thus further increasing slide review efficiency. My role was to provide technical leadership and management to guide this effort. PMA approval for this application was granted in December, 2008 after my departure.

Hard disk-file engineering.1. Mechanical reliability of head-disk systemsDisk file life is too lengthy to evaluate directly. Life estimates involved stress-tests, use of statistics, experimental design, failure analysis, materials science, and use of sensors to characterize the disk to head interface.2. Detection algorithms used in magnetic recordingBinary data used by computers must be detected in an analog signal presented by a read sensor. I participated in the development of a new detection technology using adaptive filters and maximum likelihood detection.3. Read/Write sensor researchDeveloped concepts to extend the performance of future disk file products. The work required development and use of device physical models, performance models, experimentation, tester design, product planning and coordination with researchers of diverse specialization.