企业级LLM和RAG与NVIDIA和Oracle [LRN2426]（1）.pdf

1、 Enterprise LLMs and RAG with NVIDIA and OracleLRN2426Eddie Weill,Sr.Manager,Solutions Architecture,NVIDIAShawnna DelHierro,Chief Information Officer,SoundHound AISid Padgaonkar,Senior Director,Product Management,OracleSession Agenda2Copyright 2025,Oracle and/or its affiliates|Confidential:Internal/

2、Restricted/Highly Restricted1 NVIDIA AI Enterprise Introduction2NVIDIA AI Enterprise on OCI3NVIDIA AI Enterprise on OCI with SoundHound AI3Closing and Q&ANVIDIA AI Enterprise IntroductionEnterprise Generative AI JourneyFrom insights to outcomesPerceptive AIGenerative AISingle application use casesAg

3、entic AIAI Factories creating agents at massive scaleDetecting PatternsGenerating AnswersGenerating OutcomesNVIDIA NIM Powers Generative AI FactoriesBuilding blocks of multi-modal agentic AILanguage&ReasoningRegional and Domain SpecializedVisualRAGSpeechDigital HumanHealthcareComputer Vision&Simulat

4、ionSNVIDIA NIM Optimized Inference MicroservicesRapidly deploy reliable building blocks for accelerated generative AI anywherePortable Run cloud-native microservices on OCI Platforms,maintaining security and control of data and appsEasy to Use Move fast with the latest agentic AI building blocks for

5、 reasoning,retrieval,images and more,deployed in minutes with standard APIsEnterprise Supported Gain confidence with stable APIs,quality assurance,continuous updates,security patching,and supportPerformance Optimize accuracy,latency and throughput to meet requirements with lowest TCONVIDIA NIMµ

6、servicesAI AgentsNVIDIA BlueprintsAvailable on Reference ApplicationSample DataReference CodeArchitectureCustomization ToolsOrchestration ToolsBlueprints3D Animation PipelineWeb Front EndDigital Humans for Customer Service$125B market for digital human economy by 2035BenefitsIncreases engagement and