Imagine AI
Dashboard

Danny Williams

Machine Learning Engineer · Weaviate · 10.4K followers · LinkedIn benchmark, all time (original)

Weaviate company report →
data through 2026-07-17

Report card — all time

Avg eng/post

355

Median eng/post

196

Posts/wk

0.6

Total eng.

19.5K

Trend (monthly — bars are posts, line is avg engagement per post)

Top categories

  • Industry Insight — smart avg 397.7 across 37 posts (2× typical)
  • Product Update — smart avg 287.4 across 15 posts (1.5× typical)

Top posts

  • 2025-03-24 · Industry Insight1.8K eng. (1.6K likes · 61 comments · 193 shares)

    Single-vector embeddings are so 2022. There's a quiet revolution happening within multi-vector models, and you'll never look back. Typical v

    View post →
  • 2025-10-07 · Industry Insight1.5K eng. (1.3K likes · 44 comments · 137 shares)

    I've been building RAG systems wrong this whole time. Turns out we're processing 90% irrelevant information. REFRAG fixes this massive incon

    View post →
  • 2025-01-30 · Industry Insight1.4K eng. (1.2K likes · 35 comments · 133 shares)

    You might think that DeepSeek R1 was trained the same as other top LLMs. You’d be wrong - it’s way different. Most LLMs are trained in a few

    View post →
  • 2025-05-15 · Industry Insight1.2K eng. (1.0K likes · 63 comments · 112 shares)

    I've spent months fighting with LLMs to write correct queries. This paper finally solves it. Instead of forcing LLMs to generate SQL directl

    View post →
  • 2025-09-17 · Industry Insight1.1K eng. (921 likes · 23 comments · 134 shares)

    You might think all chunking strategies are created equal. You'd be wrong - the difference is 𝘮𝘢𝘴𝘴𝘪𝘷𝘦. Chunking is arguably the most

    View post →
2024-09-052026-08-04 · Benchmark by Imagine AI · methodology at benchmark.imagineai.me/dashboard/methodology