Sr. Oracle Programmer

(Jun 2003 - May 2007)

Developed a robust and efficient matching system that integrates Court Filings, Business, and Legal Directories, enabling a comprehensive view of the current Legal Marketplace.

DATA ENGINEERING

  • Built data warehouse integration of LexisNexis Legal directory vs 80 million daily federal court filings

    • Agent-swarm parallel processing technique dramatically reduced run times

    • Integrated external address validation platform into the predictive matching algorithm

  • Stored high-confidence matches as ‘thesaurus’ entries so successive runs would automatically make match / reduced manual effort

DATA SCIENCE

  • Designed a weighted, scored confidence rating system based on daily data matching of low vs high-quality data

  • Chose training data and assigned input weights based on univariate analysis of the first 1k manually matched records

  • Built fault tolerance and machine learning into the system by storing high-confidence matches as thesaurus entries 

SOLUTION ARCHITECTURE