Traceability and Observability in Multi-Step LLM Systems - LLMOps
Offered By: LLMOps Space via YouTube
Course Description
Overview
Explore advanced techniques and best practices for implementing traceability and observability in multi-step LLM systems in this informative talk by Marc Klingen, CEO & Co-Founder at Langfuse. Gain insights into leveraging OpenTelemetry (OTel) for deep understanding of RAG systems. Delve into the complexity of multi-step LLM applications involving multiple agents and chains. Examine a detailed integration example of RAG systems, demonstrating how user input, message history, and additional memory combine to produce aggregated responses with accompanying sources. Learn about the crucial role of traces in observability and their power as an abstraction for logging and monitoring in both production and development environments. This 49-minute session, presented by LLMOps Space, a global community for LLM practitioners, offers valuable knowledge for those interested in deploying LLMs into production.
Syllabus
Traceability and Observability in Multi-Step LLM Systems | Langfuse | LLMOps
Taught by
LLMOps Space
Related Courses
Large Language Models: Application through ProductionDatabricks via edX LLMOps - LLM Bootcamp
The Full Stack via YouTube MLOps: Why DevOps Solutions Fall Short in the Machine Learning World
Linux Foundation via YouTube Quick Wins Across the Enterprise with Responsible AI
Microsoft via YouTube End-to-End AI App Development: Prompt Engineering to LLMOps
Microsoft via YouTube