Benchmark DashboardDanny Williams
Machine Learning Engineer ยท Weaviate ยท 10.4K followers ยท LinkedIn benchmark, last 12 months (original)
Weaviate company report โReport card vs prior 12 months
Avg eng/post
358
-1%Median eng/post
187
-5%Posts/wk
0.2
-75%Total eng.
4.7K
-69%Trend (monthly โ bars are posts, line is avg engagement per post)
Top categories
- Industry Insight โ smart avg 377 across 10 posts (2ร typical)
Top posts
- 2025-10-07 ยท Danny Williams ยท Industry Insight1.5K eng. (1.3K likes ยท 44 comments ยท 137 shares)
I've been building RAG systems wrong this whole time. Turns out we're processing 90% irrelevant information. REFRAG fixes this massive inconโฆ
View post โ - 2025-11-05 ยท Danny Williams ยท Industry Insight823 eng. (706 likes ยท 24 comments ยท 93 shares)
Stop predicting tokens with your LLM. Start predicting vectors. Every modern LLM, from GPT to DeepSeek, follows the same fundamental processโฆ
View post โ - 2026-02-02 ยท Danny Williams ยท Industry Insight608 eng. (524 likes ยท 17 comments ยท 67 shares)
91.3% accuracy vs 0%. Same model. Same task. The only difference: treating your prompt as code instead of text. Recursive Language Models (Rโฆ
View post โ - 2025-11-26 ยท Danny Williams ยท Industry Insight516 eng. (480 likes ยท 6 comments ยท 30 shares)
I spent the last year building production AI systems. Most developers are solving the wrong problem. Every developer building with LLMs hitsโฆ
View post โ - 2026-03-16 ยท Danny Williams ยท Industry Insight297 eng. (251 likes ยท 5 comments ยท 41 shares)
This paper will change how you build RAG systems. It asks: should you use image-based or text-based search? ๐๐ฅ๐ฃ๐๐ฃ๐๐ฅ๐ฆ: ๐ ๐ฉ๐ถ๐๐๐ฎ๏ฟฝโฆ
View post โ