This article explores how to implement key areas required for operationalizing Large Language Models (LLMs), such as safety, prompt engineering, grounding, and evaluation. It focuses on using Prompt Flow, an open-source project developed by Microsoft, to create a simple demo AI assistant for answering questions about humanitarian disasters using information from situation reports on the ReliefWeb platform.
Table of contents
A Humanitarian Crises Situation Report AI Assistant: Exploring LLMOps with Prompt FlowOperationalizing Large Language Model ApplicationsPrompt FlowWhy Prompt Flow?An AI assistant to answer questions about active humanitarian disastersSetup of the demo Prompt Flow applicationRunning the demo Prompt Flow applicationContent SafetyEntity ExtractionGetting data from the ReliefWeb APISummarizationQuestion AnsweringAttribution of Informational SourcesAutomatic Fact CheckingLLMOpsDeploymentFinal ThoughtsReferences32 Impressions