Get this course for free with Premium AI+ subscription

Workshops:

Through hands-on exercises and architectural deep-dives, participants will learn about the key differences between these approaches and when to select each one based on their specific use cases. At the end of the session, attendees will be able to implement both architectures, articulate their fundamental differences, and understand how real-world requirements drive architectural choices.
In this session, we explore the technical architecture behind creating a multimodal AI assistant using advanced RAG techniques integrated with LlamaIndex for efficient data retrieval across diverse sources. We will discuss how to address the limitations of native RAG models, including challenges with incomplete data, reasoning mismatches, and handling multimodal inputs like text, tables, and images. By leveraging LlamaIndex, along with visual language models and embedding techniques, we enable cross-modal understanding and more accurate information retrieval. Attendees will learn how to utilize LlamaIndex for structuring and indexing large datasets, combine it with LangChain-based embeddings, and implement query decomposition and fusion strategies to enhance AI performance.
This workshop gives you hands-on knowledge on using ReBAC to safeguard sensitive data in RAG pipelines. We'll start with why Authorization is critical for RAG pipelines, and how Google Zanzibar achieves this with ReBAC. Attendees will learn how to pre-filter and post-filter vector database queries with a list of authorized object IDs to improve security and efficiency.

The workshop uses Pinecone, Langchain, OpenAI, and SpiceDB.

Retrieval-augmented generation (RAG) has become increasingly popular in the latest wave of Generative AI applications, from sophisticated question-answering systems to advanced semantic search engines. This has led to an bewildering explosion of RAG methodology and research. How should we design optimal RAG systems given the breadth of design choices available?

In this workshop, we demonstrate how to quantify the performance of retrieval and generation in your RAG system using only open-source tools such as LangChain, HuggingFace, Milvus and RAGAS. Attendees will learn how to implement a complete RAG evaluation pipeline to explore design choices in a principled way, and leave with practical code examples and best practices that can be applied in real-world scenarios.

We'll build a pipeline to answer natural-language queries like "What do users say about battery life?" or "What are common complaints?" We'll first apply semantic search methods to locate relevant reviews, then apply structured generation to reliably extract structured data points like review categories, reported issues, and sentiment. We'll demonstrate how structured generation achieves this reliability by constraining LLM outputs to exact schemas with the Outlines library. You'll also learn how to extract multiple data points in a single LLM call for reliable and cost-effective analysis.
You’ll learn how to incorporate images into GraphRAG and customize graph schemas as well as search that combines visual elements. We’ll walk you through the high-level architecture and the use of associative intelligence to transform search and analytics. Notebooks that illustrate creating embeddings and creating a multimodal graph from image decomposition will be provided so you can explore how mmGraphRAG can be applied to specific domains. We’ll also leave time to discuss the implications of adding graph pattern analytics to images.

Session outline:

- Building application with contextual search capability

- Deploy Application

- Test Application - running queries

- Add LLM client and RAG searcher

- Deploy new RAG application with GPU inference

- Test RAG application

Upcoming Events: