Enterprise Insights Platform

  • This platform underpins some of Providence St. Joseph Health's most personalized and measurable internal data experiences.

    Built mostly on open source software, it helps information services scale caregiver and patient insights through complex AI powered mashups of data and text content.

    The platform originated from my personal proof of concept on a handful of desktop PCs pirated from abandoned office cubicles. It demonstrated how Apache Solr's search engine, Smartlogic's Semaphore semantic graph, and Twigkit's search application framework could improve the value of enterprise data governance and business intelligence assets by combining transactional data with textual content.

    The initiative attracted leadership buy-in and capital funding, which enabled me to operationalize the stack on 12 servers in Providence's internal cloud infrastructure. Beyond the technical bits, I also navigated the platform through vendor agreements, IS security, networking, help desk, marketing, contracting, legal, and other enterprise control processes.

  • Enterprise search platform

    Providence's internal datacenters currently host the platform stack on virtual machines in high-availability clusters for Development, QA and Production environments. By the end of 2020, they will likely be Kubernetes containers on Azure.

  • A year later, the demand for new features had outgrown the stack, so I realigned search functionalities around Lucidworks Fusion. This commercial offering of Apache Solr reduced routine platform upkeep and development time from 120 to 20 hrs/week. Application development time shrank as well after Lucidworks serendipitously acquired Twigkit.

    Fusion's end-to-end continuity between tools and resulting resource efficiency stimulated platform growth. With built-in data science, semantic knowledge graph, and analytics tools, the platform can now support a wider range of internal and public-facing applications.

    Instead of just search, it offers conversational business intelligence analytics, natural language questioning, machine-learned relevance, measurable usage metrics, and hyper-personalized results relevant to a user's context and intent.

    This aligns with Providence's overall business intelligence adoption strategy by providing one of the most compelling ways to activate the value of data and factual knowledge in an enterprise.