Lightcrest Client Case Studies

Lightcrest worked with AT&T Yellowpages.com to automate key components of their data ingest, data processing, and indexing mechanisms, reducing operational risk. Special integration with and design of Yellowpages.com's enterprise-wide issue tracking system contributed greatly to overall efficiency.

Best Buy needed to efficiently promote online catalog items without cluttering the search experience with over-engineered dashboards. By leveraging FAST ESP and FAST Impulse, Lightcrest provided a custom solution allowing Best Buy to promote specific inventory while maintaining sub-second search and a comprehensive product taxonomy.

Lightcrest worked with a top legal firm to provide search-backed information forensics and security analysis. Using custom IDS systems and packet inspection, network traffic was siphoned and indexed in real time on a 12-node cluster. Packet payloads were searched by forensics teams, allowing for the identification of key material evidence.

Lightcrest identified and executed key relevancy-improving strategies and reduced indexing time by 50%. Additionally, Lightcrest re-architected a large search cluster to provide DPS and QPS performance increases, reducing time-to-market for key customer products and services.

Lightcrest reduced operational costs and improved multi-lingual relevancy and entity extraction in the electronic discovery process for Lexis Nexis. In addition, Lightcrest provided advanced classification technologies for Early Case Assessment by leveraging machine learning techniques to reduce client man-hours by over 25%.

Lightcrest built custom connectors for Life Technologies to enable federated enterprise search across multiple in-house platforms, ultimately exposing more content of high relevancy to users across the global organization.

Lightcrest increased revenues and search relevancy for Millipore by providing product search capabilities in Korean, Japanese, and Chinese (simplified and traditional). Product hit counts for Asian market customers increased significantly, resulting in considerable revenue growth in Asian markets.

Lightcrest worked with ProQuest to identify query latency issues plaguing the search experience. I/O benchmarking and in-depth query analysis revealed optimizations and firmware upgrades that allowed for a ten-fold increase in query performance.

Lightcrest integrated custom document processing stages allowing Reed Elsevier to use Teragram content classification technology in their document processing pipeline. Lightcrest consolidated legacy Python code into a single, cohesive stage that reduced total system load by 50%.

Lightcrest built custom ingest processors for the United States Government Patent Office, allowing for more streamlined and efficient categorization and indexing of unstructured and structured data, allowing for a more comprehensive user experience.