|
You may download a PDF version of this resume here 
Keith Wiley, PhDContact
Skills
ExperienceS3 Global & Seattle Sport Sciences Inc., Dec 2017 – Oct 2023 Data Scientist Data analytics for sports performance and strategic planning. Involvement in all stages of pipeline development (data capture, analysis, visualization, reporting). Most of the following tasks were critical and central to project success and business advancement. Many were developed in direct tandem with the CEO and other company leaders. • Project owner: multi-camera deployment/configuration in professional soccer stadiums. • Critical: Created custom tools for multi-camera GUI control and recording (PySide, Tkinter, Jupyter widgets, PTP-IP APIs). • Created game-recording web-scheduler and driver (using at, similar to cron) to enable cross-timezone hands-free recordings in Europe driven from Redmond office. Other team members relied heavily on this tool. • Critical: Created user-free (i.e., automated) on-the-fly video quality verification (OpenCV) of ~25 simultaneous 4K/30fps video streams, with immediate problem alerting, visualization, web-reporting. • Created user-free, on-the-fly video transfer system to move videos from remote sites to in-office servers during recording for real-time (and post-recording) processing/analysis/reporting. • Image/video-processing toward high-level analytics: Ball movement (optical flow analysis), player pose/motion/visual-attention/skill, and team coordination (OpenCV, Pandas, lots of raw geometry). • Support of CV/ML system (in-house GPU servers). Automated data shepherding between Pandas and SQL to build hands-free, heterogeneous Python data-processing pipelines and reporting. • Project owner: Created business critical web-based reporting/publication system of player skill/tactics/strategy. • Analysis and visualization of player metrics. • Production and publication via web UI as Python-generated PDFs by a hands-free/turn-key system (my design) for consumption by high ranking soccer club owners/managers. Integris Software Inc., Oct 2016 – Dec 2017 Senior Software Engineer Distributed data pipeline development for a nascent startup's GDPR compliance product. Processing in Python. Tools included PyCharm, Jupyter, and Apache Storm. Beyond technical implementation, work involved overall product design of data manipulations and analytics toward insight and business value. Atigeo, Aug 2013 – Sep 2016 Senior Software Engineer & Data Scientist/Engineer Hadoop/Hive/Hue/Spark Python for ML and predictive analytics on AWS. Implemented a Python distributed random forest in Hive/Python streaming. ETL for feature extraction and classification. • Created a scikit-learn random forest and ensemble ML pipeline for cross-fold-validated predictive analytics and insight by feature importance exposure. ROC curves and AUC analysis/visualization of ML results. • Created a pipeline to ingest/catalog/store/analyze/visualize datasets. This project implemented an SOW whose completion was the keystone of a seven-figure contract. Expedia via Slalom Consulting placement, Dec 2012 – Jul 2013 Big Data Engineer, Consultant Part of a team building a new user feature on the web app. • Brief MongoDB project when I first started. • Hadoop/Hive on AWS, using EMR and nonEMR-Hadoop in EC2. Tasks: EC2-to-S3 data synch., Hive stand-up, AWS profiling. • Accomplishments: Hadoop 2.0/YARN EC2 deployment. Amazon's own engineers were curious about my progress. https://keithwiley.com/writing/HowToDeployHadoopYarnOnEC2.shtml Slalom Consulting, Feb 2012 – Jul 2013 Big Data Engineer, Consultant National Mobility team (mgr. Jeff Rubingh), National BI team (mgr. Kevin Gregory), developing big data processing techniques. • Focus: Hadoop MapReduce, Hive, Cloudera, Tier 3, Hadoop-on-Azure. • Topics: CRM, NY MTA, Linked-In/Twitter APIs, some OpenLayers visualization. University of Washington, Dept. of Astronomy, Feb 2010 – Jan 2012 Research Scientist IV LSST group (mgr. Andrew Connolly). Massively parallel image processing in Hadoop, e.g., image coaddition (multiple partially overlapping images registered/stacked/mosaiced). Test dataset: SDSSDB (30TB, 4-mil images), future applications to LSST (60PBs). Cluster (NSF CluE): 892 machines, 700TB storage, 3568 concurrent processes. University of Washington, Applied Physics Lab, May 2007 – Feb 2010 Software Engineer IV • Two main projects while holding this position: • Sonar Simulation Toolkit (eigenray model of underwater acoustics, mgr. Robert Goddard). Incorporated external libraries, OO design, feature development, optimization/performance-redesign, refactorization, unit-testing. • Real-time data-acquisition/FFT-proc. with low data-loss tolerances, rapid throughput, and amenability to parallelism. University of New Mexico, 1999 – 2007
Course Instructor (Jan 2007 – May 2007) – CS241, Data structures/algorithms, taught in C. The Institute for Genomic Research, Sep 1997 – Aug 1999 Software Developer C++ bioinformatics software development for DNA sequencing tools and closure analysis. Education
Personal ProjectsSample only. Please see my website for a comprehensive listing and github for a few public disseminations. Image/Acoustic Signal Processing Keith's Image Stacker: Multi-threaded (aka parallel) image stacking, Laplacian sharpening, wavelet denoising. Used by amateur astrophotographers, reviewed online and in Astronomy and Sky & Telescope. WildSpectra (collaboration: Dr. R. Haven Wiley, Biology dept, UNC-CH): Mac real-time spectrogram analyzer, used in Dr. Wiley's research lab and by researchers throughout the acoustic-biology community. Keith's iPod Photo Reader: Extracts images from iPod .ithmb image files. Implementation required reverse-engineering the image format from scratch. Data Analytics and Neuromorphic Computing Neuromorphic CM1K Emulator: A Python emulator of General Vision's CM1K neuromorphic chip, including slides presenting modeling experiments. See personal website or github for more info.
Ships in Satellite Imagery by neuromorphic chip: Neuromorphic computer vision and image classification
Leeds Butterfly Dataset by neuromorphic chip: Neuromorphic computer vision and image classification Predictive Modeling and Dynamic Websites Movie Hurl (http://moviehurl.com): A Perl-driven website of “shaky-cam” movie ratings, offering weighted averages and personalized predictions from correlated user-pair ratings. See New York Times article below.
Houseplant soil moisture monitoring system: MicroPython on ESP microcontrollers with web server, status, notifications, & analysis Android WildSpectra Mobile: Real-time scrolling spectrograms (FFT and octave-band) on Android devices. Also: real-time waveform & FFT/octave spectrum, and post-recording editing/playback and file I/O. Shead Spreet: Spread sheet for Android devices with 300,000 installs, 8500 sales, and a 4.3/5 rating. HCI Druid (PhD thesis): Vector drawing program which permits interwoven surfaces (Celtic knots, Olympic rings, etc.) and which provides an isomorphic efficient user interface.
Simulation: Artificial life, evolutionary/genetic algorithms, cellular automata, robotics, flocking (please see my website). Web Design: http://keithwiley.com, http://moviehurl.com, http://badlandswatches.com Positions, Publicity, Awards• Movie Hurl, New York Times: http://well.blogs.nytimes.com/2015/11/14/feeling-woozy-it-may-be-cyber-sickness, 2015. • Numerous interviews & articles following my book's publication (see my website for links), 2014–2015. • Advisor to the Brain Preservation Foundation, 2014–present. • Science Advisor to the Lifeboat Foundation, 2011–present. • Proceedings chair for the Computer Science at UNM Student Conference committee, 2006. • Sky & Telescope magazine. Software review: Keith's Image Stacker and Keith's Astroimager, p. 110, Aug 2004. • First place in the first International Online Artificial Life Creator's Contest, Cyberbotics Webots, khepera robot sim., Jul 1999. Graduate Research (sample)
Winter 2003–Summer 2006, Ph.D. thesis PublicationsBooks A Taxonomy and Metaphysics of Mind-Uploading. Humanity+ Press and Alautun Press, 2014. Peer Reviewed The Fallacy of Favoring Gradual Replacement Mind Uploading Over Scan-and-Copy. JoCS, 2016. Astronomy in the Cloud: Using MapReduce for Image Co-Addition. PASP, 2011. Astronomical Image Processing with Hadoop. ADASS, 2010. Parallel Distributed Image Coaddition with Hadoop. Yahoo Hadoop Summit, 2010. Representation of Interwoven Surfaces in 2½D Drawing. IEEE CG&A, 2006. Representation of Interwoven Surfaces in 2½D Drawing. CHI, 2006. Submitted or Under Review Mind Uploading and the Question of Life, the Universe, and Everything. 2015. The Fermi Paradox, Self-Replicating Probes, and the Interstellar Transportation Bandwidth. 2011. (also available on arxiv). Invited Long Exposure Webcams and Image Stacking Techniques. The Art and Science of CCD Astronomy, 2005. |
||||||||||||||||||||||||||||||||||||||||
