Andrew Ng's LinkedIn strategy, broken down
2.5M followers, 18 posts in 13 weeks, zero carousels. Here's what the data from Andrew Ng's LinkedIn actually shows operators worth copying.
We pulled 18 posts from Andrew Ng's LinkedIn over 13 weeks, along with his outbound comment history and full profile audit. The numbers are counterintuitive for someone with 2.5 million followers. If you're building pipeline through LinkedIn expertise, there's more here worth copying than most influencer breakdowns will tell you.
Andrew Ng posts 1.4 times per week, uses text-only format two-thirds of the time, makes almost no outbound comments, and runs zero carousels or video. His highest-engagement posts come from bold counter-narrative claims and personal anecdotes, not course announcements. The strategy works because his authority is established elsewhere; the posting cadence is distribution for a reputation already built. Operators without that pre-existing authority need to adjust what they borrow.
The raw numbers
18 posts in 13 weeks. That's 1.4 per week, with zero posts on Fridays, Saturdays, or Sundays. Mondays carry a third of the load (6 posts), with Tuesdays, Wednesdays, and Thursdays roughly even at 4 each.
Average reactions per post: 5,445. Average comments per post: 325. Top post (Context Hub announcement): 13,316 reactions. The "there will be no AI jobpocalypse" post: 10,883 reactions and the highest comment volume.
Media mix: 67% text-only, 33% single image, 0% carousels, 0% native video, 0% articles, 0% polls.
Outbound engagement: 31 comments in the full dataset window, spanning from 2018 to January 2026. That works out to almost nothing on a per-week basis.
Those numbers paint a specific picture. Ng is not grinding LinkedIn. He posts less than twice a week, almost never comments on other people's content, uses no format complexity, and still averages 5,000+ reactions per post. Understanding why that works, and what it doesn't tell you, is the actual value of this breakdown.
Why low volume works for him (and why it might not work for you)
Ng's LinkedIn presence is downstream of his reputation. DeepLearning.AI, Coursera, Stanford, Google Brain. The follower base came with the territory. When he posts, the feed amplifies it because the account already has mass. The posting cadence is not building the reputation; it's capitalizing on one that was built through years of external output.
The operators we audit at much smaller scales are often trying to use posting to build the reputation in the first place. That changes the math significantly. Posting 1.4 times per week at 3,000 followers means your content reaches a small enough pool that low volume will keep you invisible. Ng can afford to post infrequently because the algorithm already treats his account as high-authority. You may not have that cushion yet.
What's portable: the format choice. Text-only posts performing comparably or better than image posts tells you something about where Ng's audience is. They're reading for ideas, not stopping for visuals. If you have an audience built around your thinking, text-only with a strong opening sentence will often outperform a polished image post.
The three hook types that actually drove engagement
Ng runs five distinct hook patterns across these 18 posts: course announcements, personal anecdotes, bold thesis statements, industry observations, and tool or feature announcements. The distribution matters less than which ones generated outsized responses.
Bold thesis statements performed highest in comment volume. "There will be no AI jobpocalypse" is a single declarative sentence that contradicts a prevailing fear in Ng's audience. It got 10,883 reactions and the data shows it drove significantly more comments than the average course announcement. The pattern here is confidence plus counter-narrative. The post opens by stating the opposite of what many followers believe, which forces a response.
Personal anecdotes also outperformed expectations. The post about his daughter's birthday cake (using Gemini to design a cat-themed cake in yellow) drove 7,540 reactions. The Apple laptop post, where Ng noted the new MacBook was named "Neo," same as his son, got 6,460 reactions. These are short, low-stakes, human observations that pivot to an AI point. They work because they break the format pattern. A feed full of technical analysis gets interrupted by a seven-year-old's birthday cake.
Course announcements are 44% of his content by volume but the engagement ceiling is lower. The highest-performing course post (Spec-Driven Development with JetBrains) got 4,971 reactions, which is good, but below the bold thesis posts. The pattern suggests that what Ng's audience responds to most is his opinion and his voice, not his product catalog.
For operators trying to build pipeline: the personal anecdote plus pivot structure is the most transferable pattern here. It requires no institutional authority to execute. Pick something specific from your week that connects to what your audience cares about, state the observation plainly, then draw one inference. No four-paragraph setup required.
What the outbound engagement tells you
31 comments across years of activity, with the most recent dating to January 2026. Ng is effectively not engaging outbound on LinkedIn at all.
The profiles he has engaged with are mostly adjacent ecosystem players: the former CEO of Coursera, the CEO of Snowflake, founders in the AI/ML space. There's no pattern of commenting on posts from audiences he's trying to reach, because he's not using LinkedIn comments as a pipeline channel. He's using them occasionally to maintain relationships with peers.
This is worth being clear about: Ng's LinkedIn strategy is a broadcast strategy. Post occasionally, generate significant engagement, let the algorithm and the follower base do the distribution work. Outbound commenting would add no pipeline value for him because his pipeline comes from completely different channels.
The operators we see generating actual inbound from LinkedIn at smaller follower counts do the opposite. They comment strategically on the right posts in their audience's feed before their own posting volume would justify the reach. The comment surfaces them in front of the right people before those people would ever find their profile organically. That's the gap between what Ng does and what a 3,000-follower consultant should do. If you want a closer look at the mechanics of outbound commenting as a pipeline channel, our breakdown of Justin Welsh's LinkedIn strategy is a good counterpoint to this one.
The profile audit: where the strategy has a gap
Ng's profile scored 11 out of 100 in our audit. That's not a typo.
The headline ("DeepLearning.AI, AI Fund and AI Aspire") lists entities, not outcomes. No buyer persona, no problem statement, no conversion path. The about section is a single link to andrewng.org. No narrative, no call to action. Zero LinkedIn recommendations. No lead magnet, no booking link, no newsletter link directly accessible from the profile.
This would be a serious problem for most operators. For Ng, it's an accurate reflection of how his business works. He's not generating inbound through his LinkedIn profile. The courses sell through DeepLearning.AI's own funnel. The investment activity runs through AI Fund. The LinkedIn profile is more of a broadcast tower than a conversion asset.
If you're an operator trying to build pipeline, this is the part of Ng's approach you should not copy. Your profile has to do work his doesn't need to do. The headline should name who you help and what they get. The about section should tell a short story that ends with a clear next step. And if you want an honest read on what your current profile is actually doing (or not doing), the patterns we've documented across top LinkedIn creators gives you a baseline to compare against.
What's actually portable
Three things from Ng's content approach translate cleanly to smaller accounts:
The text-only default. Two-thirds of his posts are text only, and they outperform his image posts on a per-post basis. This matters because operators often over-invest in visual production. A clean text post with a strong opening sentence will outperform a mediocre image post most of the time.
The counter-narrative structure. "There will be no AI jobpocalypse" works because it names a fear and rejects it with confidence. For operators in any niche, the equivalent is: find the belief your audience holds that you think is wrong, and say so directly in the first sentence. No preamble, no caveats. The confidence is the hook.
The personal-to-professional pivot. Pick something specific from your life that touched your area of expertise this week. State it in one sentence. Then draw one inference. The birthday cake post is 2 sentences of personal observation followed by a point about AI capabilities. That structure is not specific to Ng's follower count or topic.
One number worth keeping in mind: his average post gets 325 comments. At 2.5 million followers, that's a 0.013% comment rate. The posts that generate outsized comment volume are the ones where he states a controversial position directly. The lesson is that density of reaction comes from the clarity and boldness of the claim, not from the visual format or the length.
Reputation compounds. The presence built over six months generates inbound without active pushing. Ng's account is the end state of that compounding. The operators worth watching for tactical guidance are the ones still in the building phase, running outbound comments alongside their posting, because that's the work that makes the eventual broadcast model possible. The patterns we track across top LinkedIn creators show what that intermediate state tends to look like.
Frequently asked
Based on our 13-week audit, Andrew Ng posts approximately 1.4 times per week, exclusively on weekdays (Monday through Thursday). He never posts on Fridays, Saturdays, or Sundays. Monday is his highest-volume day, accounting for roughly a third of all posts.
