This audio is auto-generated. Please let us know if you have feedback.
LinkedIn has revised the architecture of its feed algorithm, with the system now using more advanced artificial intelligence systems to decide what users are shown and how content is ranked in-stream.
The platform shared an overview of its updated algorithm approach on the LinkedIn Engineering blog, which details how new approaches are powering recommendations. This should lead to a more compelling, adaptive feed.
As explained by LinkedIn: “While the Feed has long been AI-powered, recent LLM advances gave us the opportunity to rethink what’s possible. That’s why we’re rolling out a new advanced ranking system, powered by LLMs and GPUs, that better understands what a post is actually about and how it relates to a member’s evolving interests and career goals.”
This audio is auto-generated. Please let us know if you have feedback.
LinkedIn has revised the architecture of its feed algorithm, with the system now using more advanced artificial intelligence systems to decide what users are shown and how content is ranked in-stream.
The platform shared an overview of its updated algorithm approach on the LinkedIn Engineering blog, which details how new approaches are powering recommendations. This should lead to a more compelling, adaptive feed.
As explained by LinkedIn: “While the Feed has long been AI-powered, recent LLM advances gave us the opportunity to rethink what’s possible. That’s why we’re rolling out a new advanced ranking system, powered by LLMs and GPUs, that better understands what a post is actually about and how it relates to a member’s evolving interests and career goals.”
LinkedIn said the new system is designed to be more adaptive to evolving user interests, as opposed to being guided by historical markers.
LinkedIn’s algorithm uses all the information each user has uploaded to guide what they’re shown in-stream, including profile info, skills, geography and the content they engage with in the app.
But previous systems have been more driven by past engagement, as opposed to more recent activity, which has prompted LinkedIn to revise the system to ensure each feed is fresh and relevant each time a user logs in.
LinkedIn said its new system has a better understanding of contextual relevance, based on LLM interpretation. This also means it can understand post context better, further improving matching.
As LinkedIn explained: “In practical terms, if you happen to be interested in ‘electrical engineering’ but engage heavily with posts about ‘small modular reactors,’ traditional keyword-based systems might miss the connection. Our LLM-based retrieval understands these topics are semantically related because the underlying language model brings world knowledge learned from its massive pre-training corpus — it knows that electrical engineers often work on power grid optimization, renewable energy integration, and the infrastructure challenges.”
This means the platform is giving a more contextual understanding of related fields and topics, which should ensure more relevance within its feed recommendations.
LinkedIn said the improved ranking system will also ensure that creators have more opportunities to reach interested audiences because the system is designed to be more aligned with evolving news, as opposed to showing users older updates.
“When industry news breaks and relevant posts start getting traction, you see them within minutes, not hours,” LinkedIn said. “When you engage with content signaling a new professional interest, subsequent Feed visits reflect that updated understanding almost immediately. The system feels responsive because it is continuously updating its understanding of both content and member interests.”
LinkedIn further notes that new members with limited engagement history will now see more relevant feed recommendations, while improved auditing models will ensure more competitive fairness and a more trustworthy feed.
Also, engagement bait is on the way out, per LinkedIn: “Over the next few months we’ll be improving our systems to reduce repetitive, click-driven posts and filter out engagement bait, so your Feed feels more relevant to your interests, and not a popularity contest.”
LinkedIn specifically points to posts that include statements like “Comment ‘Yes’ if you agree,” or posts that feature a video that has nothing to do with the associated text.
LinkedIn will also downrank recycled thought leadership posts that don’t add much in terms of substance or insight.
In practical terms, this should mean that users won’t have to put as much specific consideration into how they post to maximize viewership in the app, because the algorithm should now be able to better understand the broader context and match up content to interested users.
What will that mean for post reach? Well, it should see more experts getting more traction. Then again, as every member’s feed changes, that could also mean less overall reach for each post, but potential for more engagement and interest based on topical relevance.
It’s impossible to know until it’s been in effect for some time, but it’s worth noting that the LinkedIn feed is evolving, and that should mean that expert insight on the latest news and trends should generate more interest.
