LlamaIndex for Enterprise Knowledge Retrievalin Houston

Module 1: Introduction to LlamaIndex and Enterprise Knowledge Retrieval

• Understanding LlamaIndex and its role in Generative AI applications.
  • Core components and architecture of LlamaIndex.
  • Enterprise knowledge retrieval concepts and business applications.
  • How LlamaIndex connects enterprise data with large language models.

Module 2: Python and LlamaIndex Development Environment

• Setting up the development environment for LlamaIndex projects.
  • Python environments, package installation, and dependency management.
  • Configuring API credentials and environment variables securely.
  • Creating and running basic LlamaIndex applications.

Module 3: Enterprise Data Sources and Document Ingestion

• Connecting enterprise knowledge sources to AI applications.
  • Loading PDF files, text documents, web pages, and structured data.
  • Using document readers and data connectors.
  • Extracting document content and metadata for retrieval.

Module 4: Document Processing and Chunking Strategies

• Preparing enterprise documents for effective indexing and retrieval.
  • Cleaning and normalising text data.
  • Applying document chunking and selecting suitable chunk sizes.
  • Understanding chunk overlap, metadata extraction, and document relationships.

Module 5: Embeddings and Indexing Techniques

• Converting enterprise content into searchable representations.
  • Understanding text embeddings and semantic representations.
  • Generating embeddings for document chunks.
  • Creating and managing LlamaIndex indexes for knowledge retrieval.

Module 6: Vector Stores and Database Integration

• Integrating vector storage with enterprise knowledge applications.
  • Understanding vector databases and similarity search.
  • Connecting supported vector stores such as Chroma, Pinecone, and Weaviate.
  • Managing metadata filters, stored embeddings, and index persistence.

Module 7: Query Engines and Information Retrieval

• Retrieving relevant information from indexed enterprise content.
  • Creating query engines and configuring retrieval workflows.
  • Exploring similarity-based retrieval and top-k selection.
  • Applying metadata filtering and query transformations.

Module 8: Retrieval-Augmented Generation (RAG) with LlamaIndex

• Building RAG workflows that combine retrieved knowledge with LLM responses.
  • Connecting retrieval components to language models.
  • Constructing prompts using retrieved context.
  • Generating context-aware answers grounded in available documents.

Module 9: Advanced Retrieval and Reranking

• Improving the relevance of retrieved information for enterprise queries.
  • Exploring hybrid retrieval and query expansion concepts.
  • Applying reranking methods to prioritise relevant passages.
  • Understanding retrieval refinement and handling complex questions.

Module 10: Response Synthesis, Citations, and Source Grounding

• Generating informative responses supported by retrieved enterprise information.
  • Exploring response synthesis strategies.
  • Including source references and supporting document information.
  • Identifying unsupported answers and reducing responses that go beyond the available context.

Module 11: Evaluation, Monitoring, and Retrieval Optimisation

• Assessing the quality of enterprise knowledge retrieval applications.
  • Evaluating retrieval relevance, answer correctness, and context quality.
  • Identifying missing information, retrieval failures, and hallucination risks.
  • Optimising indexing, retrieval parameters, and response generation workflows.

Module 12: Security, Governance, and Enterprise Deployment Considerations

• Understanding the requirements for responsible enterprise knowledge retrieval.
  • Managing access controls, sensitive information, and data privacy.
  • Protecting API credentials and configuring secure data access.
  • Exploring application architecture, scalability, logging, and deployment considerations.

Module 13: Practical Project – Enterprise Knowledge Retrieval Assistant

• Building a knowledge assistant using LlamaIndex and an enterprise document collection.
  • Ingesting and indexing a sample collection of business documents.
  • Integrating embeddings, a vector store, retrieval, and an LLM.
  • Testing queries, examining retrieved sources, and evaluating response quality.
  • Documenting implementation decisions, limitations, and possible improvements through StepMerit's LlamaIndex for Enterprise Knowledge Retrieval training in Houston, United States.