SUMMARY

• Data Scientist, Machine Learning Engineer, and Quantitative Application Programmer/Analyst with extensive experience in industrial & applied mathematics, statistical inference, modeling, and simulation with applications to finance, banking, insurance, manufacturing, logistics, supply chain, and telecommunications.
• Change Point Detection in industrial sensor telemetry streams for prediction of defects to optimize Yield and Quality, Predict Maintenance, and Guide Repair
• Image Classification, Object Detection, Object Tracking, Semantic Segmentation, and Instance Segmentation for industrial process improvement
• Lemmatization, Parsing, Stemming, Tagging, Morphological Segmentation, Sentence Breaking, and Word Segmentation for industrial application of data science to text and speech
• Statistical Inference, Modeling, and SimulationFinancial Modeling, Stochastic Modeling, Statistics, and Combinatorics for high - dimensional data analysis and simulation

TECHNICAL SKILLS

Statistical Software: R, Python, MatLab/Octave, Crystal Ball, SQL, Tableau, PowerBI, OLAP

Data Science: Classification, Scoring, Clustering, Supervised Methods, Unsupervised Methods, Decision Trees, Neural Networks, Natural-Language Processing, Time-Series Decomposition

Functional Expertise: Statistics, Mathematical Modeling, Data modeling, Time-Series Analysis, Seasonal Decomposition, Machine Learning, Predictive Modeling, Analytics, Natural Language Processing, Monte Carlo Simulation, Spectral/Frequency Analysis, Digital Signal Processing

Industry Expertise: Manufacturing, Retailing, Distribution, Finance, Health Care, Energy, Utilities, Chemicals, Oil & Gas, Agribusiness, Telecommunications, Cable, Media, Entertainment, Strategic Defense

PROFESSIONAL EXPERIENCE

Data Scientist

Confidential

Responsibilities:

• Developed computer-vision techniques to identify personnel, equipment, and workflows on the oil rig. Applied numerous statistical inferential methods and mathematical transforms to pre-process streaming-video images and sensor telemetry to enhance contrast, equalize color, hue, and grayscale histograms, detect change points, and smooth statistical noise to enhance the decision discrimination capability of the on-board sensor-fusion suite.
• Global Mining Company
• Employed Multiphysics of mass and heat transport to craft the analytic data set that supported predictive operational integrity of a fine-tuned temperature-dependent solutions mining operation.
• Physical features included thermal coupling parameters for probability-distributed brine heat capacity, advection, radiative influx, sensible heat, wind-vector decomposition, saturated vapor pressure, barometric pressure, thermal emissivity, and salinity.
• Applied Digital Signal Processing techniques to Real-Time Streaming Video images of underground mine rib wall to locate and identify commercially significant geologic features in adverse conditions of low light and reduced visibility. This work greatly advanced an expensive scaled Cloud Native AutoML approach that could not discriminate features in adverse optical conditions irrespective of applied computational power.
• Developed an impactful digital image “fingerprint” using cryptographic-grade hashing algorithm that allowed discrimination of defects, isolation of fault, and guided repair for mechanical and vulcanized splices in high-speed conveyor belts under industrial-grade loads of raw mineral ore.
• Major Bank Holding Company
• Developed predictive-modeling capability for anti-fraud measures that provide compliance with Sarbanes-Oxley regulations.
• Managed project and Agile process for a major migration of a large portfolio of Excel/VBA + SQL Server management reports to MicroStrategy + AWS/RedShift at petabyte scale for numerous financial attributes including negative balances, charge backs, and income taxes.
• Led data science team for customer segmentation at rest including k-means, k-modes, and k-prototypes as well as exotic clustering of transaction telemetry streams in-flight.
• Major Auto-Service Retailer

Data Scientist

Consulting data scientist

Responsibilities:

• Primary author of management report to re-balance TraQline U.S. Automotive Market-Share data to correct for bias in design of experiment, sample selection, and invalid statistical inference.
• Quantitative programmer for machine-learning algorithms to identify and extract, transform, and load duplicate Customer First transactional point-of-sale (POS) records into Golden records in a Master Data Management (MDM) warehouse for Manthan marketing and campaign-management applications.
• Statistical analyst and quantitative programmer for analytics protocol to establish and maintain product mix and pricing for each of 400 local retail outlets throughout Western-states’ geographies.
• Bespoke analysis of variance ( Confidential ) to establish the business value of enhancements to the retailer’s Web presence that enabled online price shopping in select markets.
• Applied econometrician for time-series decomposition, Monte Carlo simulation, and multi-variate regression for wait-time densities for a portfolio of retail stores in a nine-state geographical area.
• Lead statistician for correlation analysis to Confidential operational and mathematical performance to industry and trade-group statistical-panel research.
• Major South American Telecommunications Company
• Lead data scientist to a global solutions team to prove concept and value for customer segmentation algorithms that predict credit worthiness based on pre-pay and post-pay cellular telephone service account management. United States and International patents pending.
• Major Agribusiness & Consumer Products Manufacturer

Data Scientist

Confidential

Responsibilities:

• Lead data scientist to craft point-of-sale (POS) forecasts for the portfolio of planning accounts per uniform product code (UPC) for a global consumer products and agribusiness company.
• Utilized public and bespoke weather data, satellite-imagery, and agricultural sensor data to craft feature sets for analytic data frames to optimize logistics, support commodities merchants, and schedule both calendared and on-demand, ad-hoc supply-chain tasks. Typical platforms are RESTful/JSON or XML SOAP/XMP-HTTP Web Services, R, Kafka + Zookeeper, Storm, TIBCO StreamBase and Pentaho.
• Applied multivariate regression with thirty (30) causal variables to isolate trend, seasonality, baseline, and noise. This effort comprised the analytical capability within the Siebel enterprise resource system. Exotic Naïve methods were required to model those account and product combinations that displayed intermittency or non-periodic sales behavior within sliding time windows and across epochs.
• Major Hollywood Movie Distribution Company
• Applied statistician, data scientist, and quantitative applications programmer charged with developing and implementing customer-segmentation algorithms for in-flight entertainment choices of music, movies, documentaries, sports, and news.
• Major Professional Sports League
• Developed a sigma permutation algebra to represent the League schedule in 32-dimensional space and facilitate combinatorics analysis.
• Authored permutation engine to calculate League schedule subject to constraints for travel costs, itinerary complexity, special events, venue scheduling conflicts, League rivalries, and intra- and inter-conference television contracts.
• Cable Industry Large Multi-System Operator

Data Scientist

Confidential

Responsibilities:

• Applied machine-learning algorithms to tune the deployment of throughput capacity as a function of facility total allocated IT power load and contrariwise for a portfolio of head-end, hub, and switching closets. Captured real-time monitoring data from probes and employed that telemetry to model power consumption for IT, HVAC, fire-prevention, cooling, lighting, and ambient facility load per temperate zone, season, and other dimensional drivers of outcomes.
• Employ historical weather data to build feature sets for machine-learning algorithms to support call-center operations and network platform health. This platform improved Mean Time to Detect network anomalies by nine (9) hours, effectively comprising a real-time portal into PIM/IGPM packet performance at each (X,Y,Z,t) in the network operations center. Typical platform comprises R, Kafka + Zookeeper, Storm, RESTful JSON, SOAP Web Services with Teradata data warehousing lakes.
• Applied text-mining techniques to harvest social-media chatter to measure customer satisfaction and monitor network platform health for a time period spanning a large, national network outage. Leveraged the reactive solution to architect a predictive model for proactive network platform health maintenance and customer care.
• Developed models to categorize modem hardware performance and predict Subscriber experience based on geographical and demographic factors; implemented machine-learning techniques on merged Subscriber data with public data from as a function of home size, urban population density, weather, multi-dwelling unit configuration, and network platform health.
• Prototyped machine-learning algorithms in both time- and frequency-domain to optimize network operations for video delivery.
• Designed packet diagnostic statistical capability per PIM and IGMP specifications to monitor and report on network health. Applied machine-learning algorithms to this product design to craft a balanced scorecard in real-time to report and predict network anomalies. This capability was used by the Customer Call Center to inform Customer Service Representatives of platform health per sliding time window per latitude/longitude.
• Architected data-mining and closed-loop 360-degree predictive analytics capability for an Apache Open-Source network diagnostics platform.
• Cable Industry Market Leader

Data Architect

Confidential

Responsibilities:

• Lead data architect and data scientist on an open-source complex-event processor to provide real-time analytics for multicast cable transmission. Deep learning algorithms to distill per-channel PIM and IGPM header and payload statistics over any arbitrary sliding time window and epoch. Performed exotic cluster analysis to correlate call-center volume with network platform health, with improvements in Remedy ticket tracking in excess of nine (9) hours.
• Built the extract-transform-load (ETL) and visualization capability to monitor sales results for channel partners
• Programmed a custom SharePoint end-user document portal for a national sales-delivery dashboard
• Implemented custom tools to extract data-lineage and data-provenance from ETL graphs
• Authored Oracle PL/SQL scripts to assure data quality
• Created custom error-handling scripts to identify and remedy anomalous data streams
• Programmed a custom data parsing engine to extract and transform subscriber service data for load into enterprise reporting tool.
• Designed and developed data warehousing, business intelligence, and statistical analysis tools for manufacturing, retailing, distribution, pharmaceuticals, energy, utilities, oil & gas, insurance, finance, and public health.
• CentraLytics' intellectual property was successfully acquired by a global consulting organization in June, 2014.
• Notable projects include:
• Health Analytics Software Manufacturer

Quantitative Programmer-Analyst

Confidential

Responsibilities:

• Develop real-time reporting capability for usage metrics on treatment protocols for a Confidential ( Confidential ) and its constituent Clinical Commission Groups (CCG).
• The analytic application is integrated to the electronic medical record ( Confidential ) and monitors physician interaction with the Confidential to present the medical team with real-time links to germane articles, protocols, procedures, alerts, and regulatory constraints per the application’s analytic suite based on machine-learning algorithms which comprise the organization’s evidenced-based medicine (EBM) capability that is employed to enhance outcomes.
• Global Maritime Logistics Company

Data Scientist

Confidential

Responsibilities:

• Build machine learning capability to analyze historical routing delay, payload spoilage, and property damage as a function of weather features along the ocean passage.
• National Railway
• Used bespoke weather feeds to support the scheduling function, predict fuel usage, staff crew, and predict and resolve routing issues.
• Integrate New Analytical Dimensions into a Legacy Decision Support System that provides near-real-time information retrieval to a global insurance, financial- and risk-management organization and its institutional clients.
• The system comprises an Oracle operational data store and data warehouse. Executive decision analytics are computed and presented through the ASP-COM-MTS-Oracle PL/SQL stored-procedure technology stack.
• The immediate task was to integrate a new longitudinal capability into a legacy family of financial products including fixed-income, equity, insurance, and commingled vehicles.
• The math libraries were programmed in VB6 and compiled into COM objects that were hosted on the MTS. Further math processing was coded in the PL/SQL packages or the VB6 COM, as required. New ASP pages were constructed to expose the decision metrics and analysis to the consulting actuaries and portfolio managers.
• Used weather, satellite imagery, and shipping-log data to build predictive and forensic models to characterize portfolio risk as a function of the analytic data set features.
• Global Integrated Private Investment Group
• The client manages investments in banking, transportation, oil & gas, cellular phones, home furnishings, hotels, and resort properties.
• The requirements were to create SEC-reporting and tax-compliance efficiency by integrating the business firm’s general ledger package with a Cloud-based CCH Global fx compliance package.
• The application created efficiency of nine (9) full-time equivalent (FTE) financial analysts, which resulted in savings in professional low-value-added labor of 20,000 person-hours while eliminating careless key-stroke errors and increasing accuracy and timeliness across the regulatory-compliance function. This application freed financial analysts to return to high-value-added analysis and planning tasks for the enterprise’s mergers/acquisition function.

Quantitative Programmer-Analyst

Confidential

Responsibilities:

• Confidential is an Enterprise application that provides pricing analytics for local financial analysts in approximately two-hundred (200) Confidential corporations for a Global 100 business firm.
• The firm has integrated SharePoint server and Web services throughout its global information-technology infrastructure; the Confidential suite employs Microsoft HA001230313, which provides architecture for secure Microsoft Office communication with SharePoint Server via local Access data stores.
• Authentication is provided by Active Directory services to the local Excel spreadsheet analytics package via C#.Net function libraries, which delivers the powerful, flexible, and extensible capability of Excel to the end user while ensuring Enterprise security, extensibility, and maintainability throughout the application’s value chain.
• The data Warehouse’s business-intelligence reporting capability is delivered by the Closed-Loop Executive Analytics and Reporting (CLEAR) System, which provides a centralized analytics and reporting environment for a confederated data warehouse across corporate financial transaction business.
• Applied Monte Carlo Auto-Regressive learning models to forensically reconstruct earnings statements in anticipation of a strategic merger to create an integrated telecom capability in Russia and the Commonwealth of Independent States.
• Validated assertions made by transaction’s financial advisor that undertook a fairness opinion to value the transaction in support of the tender offer.
• Developed machine-learning algorithms to support quantitative marketing analytics and customer segmentation for a portfolio of closed-loop marketing products for telephone and cable multiple system operators.
• Director of the International Data Center, the center of excellence for quantitative methods for Arthur Andersen's International Tax desk; Applied tax law, Generally Accepted Accounting Principles, accounting periods and methods, statistics, database and custom rapid-application software development to special tax-planning and compliance projects for global tax arbitrage and litigation support for Global 1000, publicly traded business firms.
• Staff financial mathematician and quantitative programmer-analyst in charge of custom rapid-application software development for fixed-income derivative trading models, pricing engines, and order-processing systems.