Sharp AI search is AlgoLab’s novel combination of information retrieval system with pre-trained transformer-based learning models. Sharp can ingest a large corpus of open domain documents of various file formats like txt, pdf, doc, ppt, json, csv, xlsx, and html that are stored in a local / remote folder, or in a document repository like SharePoint. It can crawl web content from other sources like WordPress API, direct weblinks, and online and offline videos. 

It uses Large Language Models to generate summarized response with citation deep links to page section or video time slice. The engine can be tuned for a specific corpus without the need to write fresh code.

Technology stack used: Python, Solr, Transformer models, LLM’s (OpenAI as well as Open source)