Building an Agentic Framework for Generative AI Applications

Explore a detailed conference talk from Ray Summit 2024 where Samsara's engineering team reveals their innovative approach to building an agentic framework for Generative AI applications. Dive into their RAG-centric methodology and discover how guided augmentations enable their multi-functional agent to handle complex tasks including multilingual knowledge base queries, real-time customer data processing, and visualization generation. Learn about the framework's design principles focused on maximizing customer assistance while reducing hallucinations, complete with a practical case study of their Knowledge Base chatbot implementation. Gain valuable insights into performance evaluation techniques, guardrail implementation strategies, and the extensive data pipelining capabilities enabled by Ray in their system. Through this 32-minute presentation, understand how Samsara tackles challenges in data ingestion and achieves rapid deployment in their AI platform.