Imagine AI
All companies
All time12m90d30d
originalreposts

Zilliz

AI Infrastructure ยท 25.9K followers ยท LinkedIn benchmark, last 30 days (original)

data through 2026-09-11

Momentum vs prior 30 days

Posts

9

-57%

Posts/wk

2.1

-57%

Avg engagement

4.8

smart avg / post

-43%

Momentum

not enough posts

2nd half of window vs 1st

2026-08-31 โ†’ 2026-09-30 ยท Benchmark by Imagine AI ยท methodology at benchmark.imagineai.me/dashboard/methodology

Engagement trend

Posting volume (bars) vs average engagement per post (line) โ€” page + team, original posts

Momentum over time

Change in average engagement per post vs the previous week โ€” green gaining, red losing

Category mix

Cadence

Average engagement by day posted (Pacific time)

Team contribution

Team data through 2026-07-27 โ€” employee scrapes lag this window; low numbers here mean stale data, not a quiet team.

Company page posts โ€” counted in every number on this pageExclude

What drives engagement

Smart avg engagement per post across every post in this window. Gray rows at the bottom are under the sample floor (8+ posts per category, 8+ per format) โ€” too few posts to trust the average.

By category

Event55 posts ยท low n
Product Update4.83 posts ยท low n
Industry Insight4.81 post ยท low n

By format

Video4.94 posts ยท low n
Image4.95 posts ยท low n

Best time block: 3pmโ€“6pm UTC. Timing detail is in the Cadence chart above.

Best performing posts

  • 2026-09-07 ยท Company page ยท Event13 eng. (10 likes ยท 2 comments ยท 1 shares)

    Vector search can look fine in Elasticsearch or OpenSearchโ€”until scale turns tuning problems into architecture problems. In this webinar cliโ€ฆ

    View post โ†’
  • 2026-09-04 ยท Company page ยท Event7 eng. (5 likes ยท 1 comments ยท 1 shares)

    We just wrapped up our webinar on migrating Elasticsearch and OpenSearch workloads to Milvus. If you missed it live, the recording and slideโ€ฆ

    View post โ†’
  • 2026-09-01 ยท Company page ยท Product Update6 eng. (5 likes ยท 0 comments ยท 1 shares)

    ๐—” ๐˜ƒ๐—ฒ๐—ฐ๐˜๐—ผ๐—ฟ ๐—ฑ๐—ฎ๐˜๐—ฎ๐—ฏ๐—ฎ๐˜€๐—ฒ ๐—ถ๐—ป ๐—ฝ๐—ฟ๐—ผ๐—ฑ๐˜‚๐—ฐ๐˜๐—ถ๐—ผ๐—ป ๐—ต๐—ฎ๐˜€ ๐—บ๐—ผ๐—ฟ๐—ฒ ๐˜๐—ต๐—ฎ๐—ป ๐—ผ๐—ป๐—ฒ ๐—ท๐—ผ๐—ฏ. Teams need low-latency retrieval for liveโ€ฆ

    View post โ†’
  • 2026-09-09 ยท Company page ยท Product Update4 eng. (2 likes ยท 0 comments ยท 2 shares)

    Running vector search in production requires more than fast queries. Teams also need precise access control, better observability, and infraโ€ฆ

    View post โ†’
  • 2026-09-03 ยท Company page ยท Industry Insight4 eng. (4 likes ยท 0 comments ยท 0 shares)

    HNSW is great when you need the 10 nearest neighbors. Ask for 100,000, and the bottleneck changes. ๐—Ÿ๐—ฎ๐—ฟ๐—ด๐—ฒ ๐˜๐—ผ๐—ฝ-๐—ธ ๐˜€๐—ฒ๐—ฎ๐—ฟ๐—ฐ๐—ต is becoโ€ฆ

    View post โ†’

Worst performing posts โ€” settled posts only (published โ‰ฅ7 days before the data edge)

  • Too few posts in this window to call anything the worst.

All posts in this window

Posts

0 posts found

Sort by: