SUMMARY:

• Data Scientist with a passion for solving problems and making a difference.
• Expertise with Confidential, SAS, R, Python, Jupyter Notebooks, R Studio, SPARK, Machine Learning, SQL, Netezza, Cloud, Hadoop, Oracle, SQL Server, Db2, Artificial Intelligence, Linear Programming,, Econometrics, Text Analytics, OLS Regression, Limited Dependent Variable Regression, Time - Series, Mixed Regression Methods, CHAID, Gradient Boosted Trees, Random Forest, Hypothesis Testing, Segmentation, Principal Components, Neural Networks, Confidential, Deep Learning, Ensemble Modeling Methods, Streaming Model Deployment.

EXPERIENCE:

Data Scientist

Confidential, Armonk, NY

Responsibilities:

• Processed 1.5TB of data and constructed equipment failure predictive models for a major Confidential Client using the Confidential algorithm and a Gradient Boosted Tree Model.
• Estimated the impact on sales for a small grocer when a major national chain opens a store near by using a cross-sectional time-series model.
• Identified high performing and low performing restaurants using a regression based benchmarking technique for a regional fast-food chain.
• Used Netezza to aggregate 650 Billion Records into an hourly time-series, then constructed an Confidential time-series forecast model to predict energy consumption for a regional utility company.
• Estimated the impact of weather on sales for a major motorcycle company in four key markets using a complex statistical model.
• Developed architecture for many data science related problems in many different industries. These architectures included: internet of things, real-time deployment, cognitive, machine learning, open source, api calls and batch scoring.
• Developed Relational Databases in DashDB and Netezza.
• Created Real-time scoring processes using predictive models and APIs.
• Worked extensively with Watson technology and micro-services featured in the Bluemix Cloud Platform.
• Mentored other data scientists and fostered a team environment.

Managing Consultant

Confidential, Armonk, NY

Responsibilities:

• Developed quality control routines in Confidential to quickly identify corrupted and invalid sensor data.
• Evaluated the effectiveness of existing stick-slip mitigation tools, using complex data analysis.
• Constructed series of surface to down-hole correlation metrics to measure the relationship between surface and down-hole sensors.
• Worked with drilling engineers and geophysicists to develop a proprietary unsupervised learning algorithm based on the Hilbert Transformation.

Director

Confidential, Irving, Texas

Responsibilities:

• Led a team of diverse individuals that converted an out dated and overly complex process into a SQL server/SAS process decreasing production times by more than 90%.
• Constructed Media Mix time-series models estimating the impact of media expenditures on product sales using SAS-ETS Software.
• Generated weekly Confidential reports based on media-mix models using SAS Enterprise Guide and SQL Server.
• Client facing-consulted regularly with external clients in person and over the phone.
• Managed staff, worked cooperatively on numerous interdepartmental teams.
• Wrote and presented proposals for new business.
• Developed an in-depth segmentation of client’s broadband customers based on their television viewing habits using factor analysis.
• Conducted a life stage/demographic segmentation using CHAID on client’s landline customer base to determine optimal messaging on direct mail campaigns.

Senior Manager

Confidential, Texas

Responsibilities:

• Worked with partners across the organization to enact positive change in a multi-billion-dollar publishing, internet and advertising organization.
• Managed employees, projects and vendors.
• Using a series of logistic regression models and heuristics, determined which of the 12.5 million small and medium sized businesses in the United States are profitable to contact .
• Of those profitable to contact, created several levels of logistic regression models to estimate the advertising products each business is most likely to buy.
• Using advanced analytical techniques, estimated the profit maximizing advertising product mix for each customer.
• Developed machine learning processes using SAS and SQL to update model outputs each month, ensuring timely, relevant and accurate information is always available.
• Worked on cross-functional teams with colleagues in IT, Sales, Finance and Marketing to integrate the model output into core business systems and sales rep facing tools.
• Developed reports using SAS and SQL to measure model accuracy and ensure consistent quality control.
• Created a series of linear regression models and heuristics to estimate calls counts for yellow page ads based on the ad’s size, heading, directory and placement within the heading (more than 50 Billion potential combinations).
• Mutated the regression models and heuristics into a series of SQL queries and oracle data summaries that allowed users to return lead predictions in a few seconds.
• Analyzed internet click and search detail, built a series of linear regression models to estimate internet clicks based on the position of the ad, the category and geography searched.
• Partnered with IT to create a web interface for the tool, allowing end users to estimate call counts and internet clicks via a simple, easy to use interface.
• Partnered with sales and marketing to facilitate training and the deployment of tool end users.
• Created feed-back loop with clients to morph the product into the best possible solution.
• Served as analytical support for the initiative.
• Developed logistic regression models to target businesses likely to respond to specific offers.
• Designed control and test groups using principles of experimental design to ensure the effectiveness of campaigns.
• Responsible for post-campaign reporting and evaluation.
• Managed vendors. Ensured accuracy, timeliness and cost effective delivery.