LinkedIn Recommender System Predictive ML vs LLMs
LinkedIn Recommender System Predictive ML vs LLMs

LinkedIn Recommender System Predictive ML vs LLMs

Episode publish date
August 12, 2025 5:00 PM (UTC)
Last edit date
Aug 27, 2025 1:07 AM
Last snip date
August 27, 2025 2:02 AM (GMT+1)
Last sync date
August 27, 2025 2:03 AM (GMT+1)
Show

MLOps.community

Snips
11
Warning

⚠️ Any content within the episode information, snip blocks might be updated or overwritten by Snipd in a future sync. Add your edits or additional notes outside these blocks to keep them safe.

Episode show notes

Your snips

[00:07] LLMs Reduce Feature Engineering Burden

[07:09] Mitigate LLM Latency With Distillation Or Offline Use

[09:16] Eval Criteria Stay The Same

[11:58] Prompts Replace Much Of Feature Design

[15:30] Compare LLMs Directly Before Replacing Models

[18:29] LLMs Simplify Multi-Model Pipelines

[24:21] Use LLMs For Planning; ML For Final Ranking

[27:09] LLMs Ease Cold-Start Problems

[31:49] Smaller Models Can Beat LLMs For Narrow Tasks

[34:58] Be Cautious With Agentic AI In Production

[40:02] Older Posts Reappearing On Feeds