Today LinkedIn unveiled “Galene,” a year-long effort to scale its search engine and gather “all the economic data there is in the world — to obtain the world’s first economic graph.”
One day after launching redesigned profiles, LinkedIn’s new search architecture was designed to somewhat replace Lucene, an open-source search engine library built buy the Apache Software Foundation. According to LinkedIn the company’s search team “hit a wall” and began development of Galene last year.
However, it appears that scaling searches is no longer LinkedIn’s top priority. In a company blog post, the firm announced its intentions to build the first “economic graph” and painted a picture of its complex vision.
LinkedIn’s goal, in practical terms, is to handle complex queries such as:
- Engineers with Hadoop experience in Brazil
- Data science jobs in New York in companies where my connections have worked
- Connections of Asif or Sriram who work at Google
More, from LinkedIn:
Exploring the Economic Graph: Sophisticated search functionality needs sophisticated interconnected data — provided by LinkedIn’s professional graph. Going forward, we have ambitious plans to further enrich this data by incorporating all the economic data there is in the world — to obtain the world’s first economic graph.
The Galene search architecture has been design with this long-term vision in mind. Index terms do not have to be terms actually present in the entities. They can be attributes — such as graph edges. In fact, our Member Typeahead vertical already indexes a few different kinds of edges to help in providing social signals for relevance.
In short, LinkedIn aims to serve up faster (sometimes instant), more relevant, and more complex search results across all of its products. Although Galene remains a work in progress, LinkedIn shares that it already powers a wide range of queries on the site today.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.