-Designed and implemented high throughput algorithms to reliably extract citations from unstructured text-Developed highly efficient and fast named entity extraction techniques-Developed an author disambiguation algorithm to disambiguate authors within PubMed with a high precision and recall.-Developed highly scalable clustering method based on spectral clustering to enable clustering millions of publications, grants, clinical, and patents based on their document embeddingDeveloped a publication co-citation clustering method that allows areas of science to be extracted -from a web of over 750 million citations across 25 million publications-Supported and enhanced NLP algorithms to link NIH grants with publications and patents.-Tuned search relevancy algorithms to ensure optimal results within NIH wide search tool-Implemented secure single sign-on for web applications using NIH SAML-Supported and maintained dozens of servers in a secure and highly available virtual private cloud in Amazon EC-2-Supported MongoDB replica sets and Lucene based search (Zulia) indexes with terabytes of data and hundreds of millions of records-Maintained and expanded Zulia and MongoDB based data warehouse including Grants, Publications, Clinical Trials, Patents and Drugs.

Continuation of I Cubed Analytics.

-Designed interactive data visualizations for web based analytics tool-Developed natural language processing techniques to extract and normalize domain specific named entities from millions of documents and news articles at a rate of over a thousand documents a second-Designed system to transform documents from multiple sources into Solr for searching-Designed system for deduplicating news articles with similar content-Re-engineered web platform in GWT resulting in easier to maintain code and a more responsive web applications.

-Developed a threaded and distributed content conversion system in java in conjunction with MySQL and MongoDB that can transform millions of documents a day between various XML schemas and other formats-Researched methods of automatically extracting meta data and context from PDF documents-Used GWT to generate web interfaces

-Worked individually and in small teams to design, develop, document, test, and commission real time sortation systems using Java, C++, and C in conjunction with MS SQL Server-Redesigned a reporting system using JSP