Doug Burdick Email & Phone Number

San Jose, Ca, Us

- Architect and development lead for document conversion framework productized as Smart Document Understanding in Watson Discovery. Technical lead for cross-business unit team of 12 developers and researchers. - Document conversion framework includes deep-learning components for table recognition and OCR identification with state-of-the-art performance on widely used industry and academic benchmarks.- Extensive work building knowledge graphs from petabyte-sized unstructured data sources, such as 18 years of SEC EDGAR filings for over 30,000 publicly traded companies. - Experience with NLP and entity resolution, including experience developing applications leveraging data from the financial domain.- Developed initial version of Declarative Machine Learning (DML) language used in SystemDS (https://systemdsl.apache.org), formerly SystemML, a top-level Apache project.

Mclean, Va, Us

- Developed SQL-generation tool to perform data exchange and development of an XSD importer and exporter to the system. - Developed novel data mining techniques to identify malicious applications using Cisco Netflow data. Final product operationally deployed by clients.

Madison, Wi, Us

- Supervised by Dr. Raghu Ramakrishnan and Dr. AnHai Doan. Proposed fundamental extensions to OLAP data model to support aggregation queries over imprecise and uncertain data, such as information extracted from text (IE), sensor data, and RFID tag data. - Developed language to capture user-defined domain constraints over imprecise data, and developed scalable algorithms to incorporate these constraints into query answers over disk-resident datasets.- Efforts results in 4 referred publications (VLDB 2005, VLDB 2006, VLDB 2007, VLDB Journal 2007).

- Architected and developed PoC system demonstrating feasibility of extending IBM Business Intelligence products to support imprecise and uncertain data. Resulted in patent application.

Bethesda, Md, Us

- Designed and implemented novel data cleansing algorithms to detect and correct errors in database records, and suggest completions for missing values. - Designed and implemented novel record linkage algorithms. - Efforts resulted in 10 patent applications (3 patents granted and 7 applications pending).

Doctor Of Philosophy - Phd, Computer Science

Master Of Engineering - Meng, Computer Science

Bachelor Of Science - Bs, Computer Science