LinkedIn’s relevance models are improving, with the platform incorporating more artificial intelligence-powered systems in order to broaden the scope of its interest signals and display more relevant updates.
Late last week, Erran Berger, LinkedIn’s chief technology officer, shared new insights into the platform’s evolving relevance systems, as well as how the development of AI-powered assessment is helping keep users engaged for longer through improved understanding.
With more than 1.8 million feed updates viewed per minute in the app, it’s really the only way LinkedIn can assess this kind of activity in a timely manner, and drive improved performance based on demonstrated user interest.
As explained by Berger: “With advances in generative recommenders (GR) and large-scale sequence models, we’re rethinking the foundation of how recommendations work. Newer technology allows us to understand patterns of behavior over time versus optimize for individual interactions. And if you’re on LinkedIn, that’s a crucial leap forward in how you experience the platform — because professional identity doesn’t evolve in moments, it evolves over time. And so LinkedIn needs to do the same.”
Berger said that AI-powered generative recommendation now enables LinkedIn’s system to assess each member’s actions across the platform “as part of a continuous professional journey,” as opposed to using different models for each element of the platform in isolation.
In the past, LinkedIn used specialty ranking models for each element, including the main feed, job recommendations and ads.
Now, the platform is able to use AI systems to assess all of this activity in one group, which significantly expands its correlated interest signals.
“So if activity in one part of LinkedIn can support recommendations across the platform, we don’t view that as just isolated behavior but part of a broader, continuous trajectory of what someone is trying to achieve professionally,” Berger explained. “For example, engagement with content in Feed can influence which notifications you receive, job opportunities you’re surfaced, or even which people you may want to connect with.”
This broader scope of activity should then ensure that more relevant updates are shown to each user, thus providing users with a more engaging in-app experience.
Berger also said that generative recommenders enable the platform to expand its pool of candidate posts and content, helping to better highlight opportunities across the app.
Essentially, AI-powered systems are now driving LinkedIn’s whole recommendation ecosystem, which means more results are assessed more quickly, which should drive improved relevance and engagement.
