Data science for improving healthcare in Rhode Island.

Data science and predictive analytics for healthcare improvement.

Data science for healthcare fraud detection

Main Responsibility: Provide pre/post-sales technical expertise and product knowledge in the area of advanced analytics to assist in sales efforts and partnering activities. Selected Accomplishments:• Presented at IBM Information on Demand 2013 conference on technical details and best practices on the Netezza/SPSS Modeler integration.• Provided hands-on technical support and guidance to Partners (such as SPSS Modeler, Revolution Analytics and Mathworks) for integrating their analytic solutions with the Netezza database appliance. This support resulted in: the inclusion of at least four major new features in various SPSS Modeler releases; multiple releases of Revolution Analytics for Netezza and the ability to use MATLAB code on the Netezza database appliance.• Provided pre/post-sales support for 40+ client engagements between Netezza and various Partners (such as SPSS, Mathworks and Revolution Analytics). The support included presentations, demos and POCs demonstrating the value of the integrated solutions. This support led to sales in at least six of the pre-sales engagements and a greater understanding and faster run-times in the post-sales engagements.• Developed and presented a webinar to a technical audience of 100+ participants demonstrating the features and functionality of the Netezza/SPSS Modeler integrated solution.Detailed Responsibilities:• Provide assistance and guidance to Partners integrating their analytic solutions into the Netezza database appliance• Assist sales efforts by participating in Proof of Concepts and demos for clients• Assist in training Partners, Clients and Users on proper and efficient use of our analytic products• Write white papers and deliver webinars detailing best practices and how-tos for various integrated Partner solutions• Support technical roadmaps and product requirements for our analytic Partners and Customers• Provide appropriate tools, techniques and lab systems to enable Partner development

Main Responsibility: Design and implement enterprise fraud management/detection solutions for clientsSelected Accomplishments:· Participated in the design and initial implementation of the ACH Fraud Detection packaged solution for Memento's clients. This packaged solution allows for quicker and easier client site implementation, thereby allowing the client to start monitoring and detecting ACH fraud sooner.· Implemented fraud detection solution at Client's site, allowing them to investigate alerts more efficiently and detect/prevent a greater amount of fraud.Responsibilities:· Consult with clients to understand thier business requirements and fraud detection priorities· Translate their business requirements to technical specifications for implementation· Perform application hardware sizing as required· Install and configure Memento Security software including MS SQL, Opalis and Sharepoint.· Work with client IT organization to establish ETL infrastructure for automated daily consumption of data files· Perform ETL on client's data for use in creating and implementing fraud detection scenarios· Apply mathematical, statistical and analytical expertise to develop fraud detection scenarios· Establish infrastructure for automated daily fraud alert publishing · Configure Sharepoint-based UI for client to efficiently investigate thier fraud alerts· Consult with clients to customize UI display for data fields and management reports· Assist in training clients on proper and efficient use of our products

Main Responsibility: Predictive modeling for credit/debit card fraud detection.Selected Accomplishments:· Developed predictive Debit card fraud detection model for major client that caught 45% of the fraud accounts, by the third fraud transaction, at a 15:1 false positive ratio, saving the client 40% of their potential losses.· Developed and implemented new methodologies for the model development process which increased model performance by approximately 3% absolute.Responsibilities:·Develop predictive models for credit/debit card fraud detection·Mine and analyze data for new fraudulent patterns/trends·Create and develop predictive variables based on client transaction patterns·Write fraud analysis reports and model performance reports for clients·Develop new methodologies and processes to improve model performance·Utilize ETL tools for data preparation

Main Responsibility: Speech recognition software and algorithm development for Computer Aided Medical Transcription.Selected Accomplishments:· Enhanced a transcription filtering mechanism thereby increasing annual revenue by 1% absolute.Responsibilities:·Develop Linux-based utilities, in perl and bash, for analyzing data·Present findings at group meetings·Improve transcription filtering mechanism thereby increasing annual revenue·Assist clients with any support issues that may arise·Recommend process/technology improvements

Main Responsibility: Predictive modeling for credit/debit card fraud detection.Selected Accomplishments:· Developed predictive Debit card fraud detection model for our client that caught 40% of the fraud accounts, by the third fraud transaction, at a 20:1 false positive ratio, saving the client 48% of their potential losses.· Through data mining and analysis, discovered new fraudulent patterns in clients’ data, which allowed the creation of a new suite of highly predictive variables.· Developed and implemented new methodologies to reduce model development time, without sacrificing quality, by 10 business days overall (12.5% of the original project cycle).Responsibilities:·Develop predictive models for credit/debit card fraud detection·Develop utilities for data analysis·Utilize ETL tools for data preparation·Mine client’s data for fraud transaction patterns·Develop predictive variables based on client transaction patterns·Write fraud analysis reports and model performance reports for clients·Assist clients in implementing risk management strategies·Develop new methodologies and processes to reduce model development time

Main Responsibility: Software development for real-time medical devices.Selected Accomplishments:· Designed, developed and implemented, from getting the data to displaying the data, a GUI for mini-trend data variable display on patient monitors.· Enhanced non-invasive blood pressure measurement system to allow for a more accurate measurement.Responsibilities:· Develop real-time medical instrumentation software on a UNIX platform· Develop real-time GUI applications for patient monitors· Design real-time 3rd party device interfaces for patient monitors· Develop real-time embedded system applications for patient monitors· Develop real-time signal processing algorithms for patient monitors· Develop requirement specifications for above applications· Write and perform verification and validation tests· Work with customers to determine product development