LlamaIndex Newsletter 2024-06-25

Hello to All Llama Lovers!🦙

Welcome to this week’s issue of the LlamaIndex newsletter! This edition is packed with thrilling updates, comprehensive guides, and detailed tutorials to help you gain a deeper understanding of our tools.

🤩 The highlights:

• CrewAI Multi-Agent Integration: Integrated with CrewAI to enhance task-solving with specialized agent crews and LlamaIndex integrations. Notebook, Tweet.
• MistralAI Fine-Tuning API Integration: Enhance model training and performance monitoring with our new integration of MistralAI’s Fine-Tuning API. Notebook, Tweet.

✨ Feature Releases and Enhancements:

1. We have launched a Multi-Agent integration with CrewAI to build a crew of specialized agents that collaboratively solve tasks. Enhance these agents with external knowledge and third-party tools through easy integrations with LlamaIndex, including advanced RAG query engines and tools from LlamaHub. Notebook, Tweet.
2. We have integrated MistralAI’s Fine-Tuning API to create and synthesize training and evaluation datasets, assess model after fine-tuning, and monitor performance metrics with RAGAS and Weights & Biases. Notebook, Tweet.

💡 Demos:

• RAGapp by **Marcus Schiesser** simplifies Agentic RAG in enterprise settings with functionalities akin to using GPTs by OpenAI. The latest version includes a code interpreter and a tool to call any OpenAPI, all built using LlamaIndex.

🗺️ Guides:

• Guide to Multi-Document Agentic RAG Using LightningAI: Jay Shah’s template that enables you to set up a multi-document agent for search and summarization across research notebooks. This out-of-the-box solution, integrated with Streamlit, allows for full visualization and is part of LightningAI’s suite of tools for developing and sharing ML and genAI native apps.
• Guide to Making a Serverless RAG Chatbot: Azure’s quick start repository for creating a serverless RAG chatbot using LlamaIndex and AzureOpenAI.
• Guide to Building an Agent in LlamaIndex: Our comprehensive guide which covers building a basic agent, using local models, adding RAG, enhancing retrieval with LlamaParse, and developing custom tools.

✍️ Tutorials:

• JinoRohit’s tutorial on using a LlamaIndex pipeline with MLflow for systematic tracking and tuning of RAG parameters, enhancing answer accuracy through precise evaluation metrics and datasets.
• Hanane Dupouy’s tutorial demonstrates how to apply CRAG (Corrective RAG) for financial analysis using LlamaIndex’s CRAG LlamaPack. This technique assesses retrieval quality and supplements the knowledge base with web searches to ensure contextual accuracy and relevance.
• Soham’s tutorial to create an agent that automates GitHub commits using Composio and LlamaIndex Tools.
• Aruna Withanage’s tutorial on creating custom text-to-SQL pipelines using LlamaIndex’s DAG capabilities.