37:11
Root Talks: Lauri Pulkka, Newsec
30:10
Root Talks: Tommi Rasinmรคki, Gofore
37:11
OLD: Root Talks: Lauri Pulkka, Newsec
43:27
Root Talks: Henri Viertolahti, Gosta Labs
04:51
Testimonial Gofore
04:19
Testimonial Newsec
03:38
Testimonial Gosta Labs
04:58
๐ ๐ถ๐๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ถ๐ผ๐ป๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ ๐๐๐ฑ๐ด๐ฒ๐ - #๐ต: Judges are only for pre-production
04:43
๐ ๐ถ๐๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ถ๐ผ๐ป๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ ๐๐๐ฑ๐ด๐ฒ๐ - #๐ญ๐ฌ: You always know what to measure from the get-go
41:56
Top-10 Misconceptions About LLM Judges in Production
38:54
10 Critical LLM Blunders - Detect and Fix with LLM Judges
02:20
๐ ๐ถ๐๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ถ๐ผ๐ป๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ ๐๐๐ฑ๐ด๐ฒ๐ - #๐ด: You Always need human in the loop
02:28
๐ ๐ถ๐๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ถ๐ผ๐ป๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ ๐๐๐ฑ๐ด๐ฒ๐ - #๐ณ: Judges make sense only in A-B comparisons
03:17
๐ ๐ถ๐๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ถ๐ผ๐ป๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ ๐๐๐ฑ๐ด๐ฒ๐ - #๐ฒ: Traditional NLP metrics are enough
04:15
๐ ๐ถ๐๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ถ๐ผ๐ป๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ ๐๐๐ฑ๐ด๐ฒ๐ - #๐ฑ: Better models need no judges
02:05
๐ ๐ถ๐๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ถ๐ผ๐ป๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ ๐๐๐ฑ๐ด๐ฒ๐ - #๐ฏ: We know the risks and Judge is an overkill
03:10
๐ ๐ถ๐๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ถ๐ผ๐ป๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ ๐๐๐ฑ๐ด๐ฒ๐ - #๐ฐ: Judges are there just to block unwanted responses
02:54
๐ ๐ถ๐๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ถ๐ผ๐ป๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ ๐๐๐ฑ๐ด๐ฒ๐ - #๐ฎ: Judge models need no management
01:52
๐ ๐ถ๐๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐๐ถ๐ผ๐ป๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ ๐๐๐ฑ๐ด๐ฒ๐ - #๐ญ: Anyone can be a judge
02:52
Root Signals - 3 min demo
X
About
Welcome to the Root Signals YouTube Channel, where we empower developers to create, optimize, and embed LLM evaluators for continuous monitoring of LLM automations in production environments. Our End-to-End Evaluation Platform ensures that development teams can deliver reliable, measurable, and auditable LLM automations at scale. Subscribe to explore insights, tutorials, and discussions on making your AI systems more effective and accountable. Join us in advancing the frontier of responsible AI development.
Tags