From catching up on trending news, getting updates from your network, or following a thought leader from our Influencer program – the professional information and insights in the LinkedIn Feed have become central to our member experience. The LinkedIn feed is the sorted list of updates displayed to our members when they log-in to or use the mobile app.

Specifically, the feed serves our hundreds of millions of members by:

  • Showing updates and shares from connections, trending news and sharing insights from industry leaders.
  • Displaying content relevant to their profession and career, enabling them to stay informed and improve their career prospects.
  • Creating an opportunity to establish their professional brand by sharing and creating professional content, including long form posts.
  • In addition to helping millions of people establish a professional reputation, staying informed, and staying in touch, the feed also exposes members to relevant job opportunities and opportunities to grow their network via jobs recommendation and "People You May Know" updates.

Feed relevance is the task of evaluating the relevance of updates to different members, and more generally ranking the updates in a personalized feed format. The task of evaluating relevance and ranking updates is similar to classic recommendation systems such as movie recommendations, with a few twists:

  1. The LinkedIn feed has a heterogenous inventory of updates: articles shared by connected members, jobs recommendations, news recommendations, suggestions to connect to members (“People You May Know”), news stories mentioning companies that the member is following, etc.
  2. Many members interact heavily with the feed, scrolling down several times. It is very important to display a feed that allow the member to have engaging interactive sessions, rather than focus only on finding the most relevant updates and include them at the top.
  3. The feedback that we collect is based on members’ implicit actions, rather than explicit actions such as assigning 1-5 stars as in classic movie recommendation. Implicit actions that we track include clicks, likes, comments, shares, time viewed, etc.
  4. LinkedIn’s social graph is an important component of the relevance modeling. Members often find updates shared (or liked, or commented on) by their connections to be more relevant than otherwise.
  5. LinkedIn’s members have an incentive to complete their profiles on the site as it represents their professional identity. As a result, LinkedIn has access to high quality data representing members’ interests, skills, and background.

The LinkedIn feed relevance system includes multiple sources called First Pass Rankers (FPRs) that create a preliminary ranking of their inventories based on predicted relevance to the feed viewer. Examples of FPRs are jobs recommendations, news article recommendations, updates or shares from your connections, recommendations for new connections (PYMK), and sponsored updates (SUs). The FPRs score their respective inventory of updates with respect to the viewer, and output the top-k updates to the Second Pass Ranker (SPR) that combines the output of all FPRs and creates a single personalized ranked list. After the second pass ranking, the ranked list is passed to the reranker stage that modifies the output of the SPR and creates the final feed that is sent to the appropriate front-end.

The feed relevance team is responsible for the following efforts:

  • Offline infrastructure: developing tools (often on-top of Hadoop or Spark framework) to process big-data, create training data, train models, evaluate models.
  • Modeling: trying different machine learning models, training frameworks, response data, etc.
  • Online infrastructure: develop software that serves results to front-end in real time.