Data Scientist

ROBERT L McPHERSON

** ******* ***., ********, ** 06902

203-***-**** Office, 203-***-**** Mobile

******.*.*********@******.*******.***

www.linkedin.com/in/robertmcpherson/

PROFESSIONAL EXPERIENCE

W. R. BERKLEY CORPORATION – November 2014 to Present

Leader, Corporate Analytics, Greenwich, CT – November 2014 to Present

Lead analytics for all 50+ insurance company within W. R. Berkley Holding company.

Established center of excellence practice to provide infrastructure, consulting, and analytics support within all the companies. Directly managed a team of five senior level data scientists. Also responsible for the predictive modeling and analytics activities of more than 50 advanced data scientists, as well as hundreds of “citizen data scientists” and business analysts embedded within the Berkley companies in the U.S., Canada, Europe, and Asia.

●Increased predictive models implemented by the companies from fewer than 30 to more than 55

●Led development of microservices infrastructure to host and share predictive models, resulting in rapid deployment of complex AI based models in days, rather than months

●Provided data science training and development to the Berkley companies, contributing to a more than 60% increase in available data science talent

●Established enterprise data lake, enabling first ever cross-company integration of key business lines

●Hosted multiple Advanced Analytics Summits for the analysts and data scientists from our companies around the world

●Planned, organized, and delivered first ever Citizen Data Science training, resulting in new predictive models within the same year

NATIONWIDE – December 1985 to November 2014

Director, P&C Actuarial Research Analytics, Columbus, OH – April 2007 to November 2014

Directed a team of five data scientists responsible for advanced analytics, predictive modeling, data integration, and reporting. Developed predictive models to explain shifts in modeled loss exposure. Performed outlier detection utilizing algorithms based on predictive modeling. Created reproducible research documents for predictive modeling, GIS analysis, data quality assessments, statistical outlier detection, and other analytical processes. Provided grid computing expertise and big data support related to catastrophe modeling for all P&C operations. Developed automated ETL data processes for large analytical data sets utilizing big data platforms. Utilized variety of data science and analytical tools, including Hadoop, MapReduce, Hive, SQL, R, Python, SAS, and Java.

●Contributed to billion dollar reductions in catastrophe exposure

●Managed large projects to construct high performance virtual computing grids, produce insights through big data analytics, and build the P&C Risk Management Data Mart

●Introduced Agile and other project management techniques to the department

●Implemented code versioning and other IT related best practices

Product Director, Research & Analysis, Columbus, OH – September 2004 to April 2007

Managed team of 9 data scientists and business analysts to conduct product and underwriting research, and develop business intelligence processes for all personal lines property products. Developed automated analytical, pricing, and reporting applications in a variety of languages, including Java, HTML, VB/VBA, R, SAS, and SQL.

●Developed predictive model for selecting policies to review for home inspection program, saving more than $5 million in inspection related expenses

●Managed project to build automated policy rating tools in Java, enabling roll-out of innovative, by-peril, homeowner rating plan to millions of customers

●Created company’s first ever database and reporting processes for insurance quote data, and performed advanced statistical analysis to uncover price elasticity relationships

●Designed automated econometric analysis applications to forecast underwriting cycles

Marketing Director, Columbus, OH – March 1999 to September 2004

Managed research activities related to all Nationwide Financial (NF) products. Partnered with senior management to explore and craft future strategic direction. Uncovered strategically valuable insights by utilizing advanced quantitative methods. These included brand mapping, predictive market share models, and war gaming based on Markov Chain methods.

●Lab-tested new strategies through advanced modeling, combining predictive modeling with game theory methods

●Successfully developed brand positioning for new mortgage operation utilizing advanced customer value analysis methods and correspondence analysis; operation grew to become Nationwide Bank

●Created and managed company’s first competitive intelligence unit and war room

●Identified potential targets among NF's competitors, and specified strategic actions exploiting key vulnerabilities

Sales Technical Specialist, Huntington, WV – Dec 1997 to Mar 1999

Motivateed and trained agencies in the Charleston and Huntington sales districts. Districts achieved awards for life and auto insurance sales performance. Conducted predictive modeling research resulting in important discoveries regarding the relative effectiveness of various sales processes.

Underwriting Portfolio Manager, Martinsburg, WV – Oct 1996 to Dec 1997

Performed predictive econometric time series analysis of underwriting cycles, which allowed us to avoid a costly re-underwriting effort. Advised and trained agencies on underwriting issues and audited agency files for underwriting activities. Conducted portfolio analysis on multi-million dollar P&C portfolio. Achieved the best overall portfolio profitability and growth results in the state operation.

Claims Market Manager, Martinsburg, WV – Feb 1996 to Oct 1996

Managed 4 claims offices over large territory with responsibility for 23 employees, $2,000,000 expense budget, and claims paying activities of 40 agents for personal and commercial lines claims operations.

Various Positions & Manager Roles: Underwriting, Sales, and Marketing, Dec 1985 to Feb 1996

EDUCATION AND PROFESSIONAL DESIGNATIONS

●Masters in IT & Information Management, Harvard University, Cambridge, MA, 2013

●Master of Business Administration (MBA), University of Phoenix, 2000

●Bachelor of Science in Marketing/Business Administration, University of Akron, 1985

●Chartered Property & Casualty Underwriter (CPCU), 1996

●Chartered Life Underwriter (CLU) & Chartered Financial Consultant (ChFC), 1991

●Project Management Professional (PMP), 2007

PROFESSIONAL ACTIVITIES

●Speaker, Casualty Actuaries of New England Spring Conference, Foxwoods Casino, CT

●Speaker, Predictive Modeling, Insurance Analytics Conference, Chicago, IL

●Guest Lecturer, Security Analysis, Harvard University, Cambridge, MA

●Speaker, Big Data Analytics, Shakacon Security Conference, Honolulu, HI

●Co-author, Information Security Analytics: Finding Security Insights, Patterns, and Anomalies in Big Data, Syngress Publishing

●Textbook reviewer: Insurance Marketing, LOMA Institute, Atlanta, Georgia

TECHNICAL SKILLS

R, Python, SQL, Java, Hadoop, Hive, Spark, AWS, VB/VBA, HTML, XML, JavaScript, JSON, Machine Learning, AI, H2O, Distributed Random Forest (DRF), Support Vector Machine (SVM), Generalized Linear Model (GLM), Deep Learning, Regularization, LASSO, Elastic Net, Ridge Regression, Text Mining, Word2Vec, Doc2Vec, Micro Services, Containers, Docker, Univariate and Multivariate Time Series, Bayesian Vector Autoregression (BVAR), Ordinal Lasso, ARIMA, Econometric Forecasting, Simulation Modeling, Reproducible Research, LaTeX

Contact this candidate