TECHNICAL SKILLS:

Operating Systems/Workstations: Pyramid DC/OSX, HP/UX, Windows NT / 2000 / XP / 7 / 10, Solaris, Red Hat Linux, Mandrake Linux, Ubuntu, SuSE

Languages/Packages: UNIX/C/C++, Java, Perl, R, MatLab Script, SQL, UNIX Shell Script, HTML, JavaScript

Services: BEA Tuxedo, gSOAP, Rogue Wave Hydra, Xerces - C++, openSSL

Database: Sybase, Oracle, OpenINGRES, INGRES II, SqLite3, MySQL, Redshift, PostgreSQL

Big Data: Spark 2.2.0, SparkR, H2O, MxNet, DataBricks

Other Tools: Office 97 / 2000 / XP / 2010, Visio, PVCS, CVS, SVN, GIT, Gerrit, Rational ClearCase, Visual Studio .Net, NetBeans, Eclipse, Matlab, Jira/Confluence, Rstudio, RShiny

Development Methodologies: Waterfall, Agile/Scrum

PROFESSIONAL EXPERIENCE:

Confidential, San Diego, CA

Engineer/Scientist

Responsibilities:

• Machine Learning and Algorithm Design
• Developed a custom classification technique based upon random fields and phase transitions and a custom deep learning technique which combines an extreme learning machine with deep belief networks for a sales application which predicts future consumer printer and ink sales.
• Developed a hybrid classification algorithm whereby the signal produced by discrete data samplings from mammography images are encoded using a discrete wavelet transform before performing a classification task using a decision tree.
• Developed a probabilistic algorithm utilizing the theory of random geometric graphs and random clusters applied to the classification problem of detecting outliers in corporate network traffic.
• Developed a probabilistic neural network anomaly detector for detecting anomalies in corporate network traffic.
• Developed a probabilistic social network analysis paradigm utilizing the theory of random geometric graphs and random clusters for the problem of analyzing social network connectivity and clustering.
• Performed statistical analyses of pharmaceuticals for the purposes of determining factors leading to a switch between drug treatments or escalation of treatments between different categories/classes of drugs.
• Wrote C/C++ implementations of TCP/IP socket - level software for a rules-based, intelligent application which addresses distributed, denial-of-service attacks.
• Wrote MatLab, C/C++, Java, Perl and R implementations of statistical, probabilistic and machine learning algorithms.
• Met with prospective clients within the business domains of cyber security, software and life sciences to foster potential working relationships, gain knowledge of their particular pains and to discover where I might provide some additional relief.
• Discussed implementation of different data sciences methods to address the problems of intellectual property theft, distributed denial-of-service attacks, pharmaceutical data mining and anti-prescription drug fraud.
• Developed and prototyped machine learning algorithms for intellectual property theft.

Confidential, Irvine, CA

Sr. Engineer/Scientist

Responsibilities:

• Machine Learning and Algorithm Design
• Developed a custom deep learning technique which combines an extreme learning machine with deep belief networks for a hierarchical classifier which predicts future service for Confidential printer and facsimile maintenance.
• Developed multi-dimensional data flattening schemes for increasing throughput and analysis using advanced statistical distribution estimation techniques.
• Developed feature selection schemes for data reduction utilizing advanced probability and statistical techniques based upon the Random Cluster Model.
• Developed missing data estimation techniques using advanced probability and statistical techniques based upon the Markov property and stationary time series.
• Developed natural language processing algorithms for finding a list of service repair actions which correspond to Confidential printer and facsimile issue events.
• Systems Design and Prototyping
• Prototyped deep learning methods to analyze Confidential printer and facsimile issue event and usage data to predict when a device would need to be serviced.

Confidential, San Diego, CA

Engineer/Scientist

Responsibilities:

• Mathematical Learning and Algorithm Design
• Designed artificial neural network algorithms utilizing advanced probability and statistics for analog sensor response modeling.
• Systems Design and Prototyping
• Prototyped the model-based development of portions of a human ventilator system, including gathering and documenting requirements, estimating the development effort in terms of time and cost, architecting the sensor data-acquisition model and ventilator control model, producing a statement of work, dividing the tasks into a logical work-flow and producing design documentation.
• Developed a MatLab model of a patient ventilator system in Linux/C/C++ and MatLab/Simulink, including continuous digital pulse generation for driving the ventilator pneumatics, processing of digital signals for conversion to human-readable, patient vital characteristics, along with ventilator control systems designs.

Confidential, Seal Beach, CA

Engineer / Data Scientist

Responsibilities:

• Machine Learning and Algorithm Design
• Developed supervised learning and natural language processing algorithms utilizing advanced probability theory and latent dirichlet allocation for Boeing 737 systems document classification.
• Developed artificial neural network algorithms utilizing advanced probability and statistics for prediction of Boeing 737 service requests.
• Developed a least squares algorithm that gives a linear estimate of measured data for accuracy tests against estimates produced by data smoothing using a Kalman filter.
• Developed a Kalman filter algorithm for data smoothing which used the least squares algorithm in order to produce an initial state estimate of a linear system.
• Developed a QR algorithm which factored a given matrix into the product of an orthogonal matrix and an upper-triangular matrix for use in finding the eigenvalues of the original matrix in determining the stability of a linear system.
• Developed an algorithm based upon topology and convex sets which was used to project an error ellipse from the perspective of an F-15 aircraft in flight onto the surface containing the target of an unguided munition.
• Developed a statistical testing methodology from the theory of design of experiments used for fault detection, error prediction and analysis in Future Combat Systems ( Confidential ) Battle Command hardware and software systems components.
• Data Analysis
• Performed analyses of commercial flight data to determine probable cause of failures in Boeing 737 aircraft systems.
• Performed analyses of target imagery data using Matlab to test the efficacy of a topological and convex sets technique to project a targeting error ellipse from the perspective of a moving body onto a stationary target .
• Performed analyses of flight test data using a command line flight simulator to perform a trade-off study of implementing a linear least squares trajectory estimate versus a Kalman filter trajectory estimate .
• Internal Client Relationship Management
• Attended weekly meetings with Boeing 737 client representatives to determine latest reported troubles with the fleet and to mitigate associated risks.
• Produced data visualizations in R or Tableau to detail airline operator's issues and to make a determination if reported problems were systemic or isolated.
• Systems Design
• Built a system consisting of Linux guest virtual machines to host a MySQL RDBMS with a DB and tables replicated from a live, production Teradata DB and an application and web server for Rstudio/Shiny server and to perform machine learning algorithm development and visualizations.
• Software Engineering
• Developed Linux/C/C++, Perl, Java and R code of statistical natural language processing and machine learning algorithms to determine probable cause of failures in Boeing 737 aircraft systems.
• Developed Java, UNIX/C/C++, Matlab script, Perl and Korn shell scripts used for F-15 munitions modeling, image analysis and processing, fault detection and statistical error analysis on myriad defense programs.
• Developed Linux/C++ code to perform text analysis and topic classification of 737 systems logbook data.
• Developed Java code to perform hybrid, K-means classification of 737 flight systems data to determine which correlated factors cause systems failures.

Confidential, San Diego, CA

Sr. Engineer/Scientist

Responsibilities:

• Mathematical Algorithm Design
• Designed a Poisson-distributed random number generation algorithm for generation of unique transaction IDs for each SOAP request/response message.
• Software Engineering
• Wrote Linux/C++ web-services against gSoap and Rogue Wave Hydra SOAP servers, along with Xerces-C++ for XML validation, for an intelligent, rules-based electronic medical records application, which combines data from medical and pharmaceutical claims records to process claims and raise alerts about potentially harmful drug-drug interactions or drug-disease adverse effects.