Sajari is a real time cloud based search and recommendation engine. It is designed to work with very large queries, like documents, peoples profiles, etc. Uses include HR management, patent discovery, journal searching, tailored news, etc.
The technology supports locations, entities, query boosting and many more cutting edge features. It’s also a self learning engine and can adapt to the users desired context in real time.
Sajari came about from the frustration of being forced to search with keywords. That sounds crazy, but there are many applications where a search is actually aimed at finding related or similar information and the searcher is often in possession of very rich contextual information (e.g. documents) that they would like to use as a search query. The standard process was to think about those many documents you would like to use as a query, summarize them in 3-4 words, throw them away, then search and hope for the best.
Summarizing complex documents in 3-4 words is very difficult. This was particularly a problem for academic literature searching and intellectual property searches, but since we started working on the problem we saw it actually occurs everywhere. They discussed it a few too many times and ended up building it.