Searchplex

Searchplex was hired by CuratedAI, a LegalTech startup, to develop a retrieval system for legal document searches. The goal was to add keyword search to semantic search, make the search faster, and improve the document search algorithm. Searchplex was chosen for their cultural fit, cost value, and aligned company values. One consultant from Searchplex was assigned to this project. The retrieval system was built to handle multiple languages and provide real-time search results with minimal latency. It was designed to continuously update with new legislation and case studies, and special emphasis was placed on privacy for user-uploaded documents. The system initially supported only English queries and was limited to European legislation. The consultant implemented a Qdrant vector database hosted on a virtual machine using Docker containers to facilitate real-time search. Document updates were done manually through scraping and ingestion, and a Python script was used to split the documents into meaningful sections.