Gen AI - RAG Application Development using LlamaIndex

Name: Gen AI - RAG Application Development using LlamaIndex
Rating: 4.2 (86 reviews)

Learn LlamaIndex to Develop RAG Applications using Open AI GPT, Gemini LLM and Vector Databases

Created byManas Dasgupta

Last updated 8/2024

English

What you'll learn

Fundamentals of LLM RAG Application Development
Using Open AI GPT API to develop RAG Applications
Prompt Engineering - Write Optimized Prompts for your RAG Application
Using LlamaIndex Query Engines, Retrievers and Query Pipelines
Building Conversational Memory
Using Data Connectors
Building Smart Agents and Tools
Language Embeddings and Vector Databases
Working with SQL Databases
Working with Structured Data and Dataframes in RAGs
Convert your LlamaIndex RAG as a FAST API

Course content

2 sections • 21 lectures • 7h 37m total length

Course Introduction14:25
Explore the foundations of retrieval augmented generation with LlamaIndex, large language models, and vector databases, and learn to build rag applications with pipelines, prompts, and Streamlit interfaces.
Introduction to LLMs32:31
Explore large language models, their transformer architectures and generative capabilities, including text generation and summarization, plus retrieval augmented generation with enterprise data.
Introduction to LlamaIndex39:52
Introduction to Prompts17:33
Prompts - Advanced18:53
Explore advanced prompting techniques, including chain of thought, self-consistency, and prompt chaining, with hands-on OpenAI playground demonstrations and Rag data considerations for LlamaIndex.
Setup your Development Environment47:48
Set up your llama index rag environment by creating a project folder, configuring a virtual environment, installing llama index and dependencies, and loading OpenAI or Gemini keys via dot env.
Your first LlamaIndex Program32:20
Create your first llama index program that ingests pdf documents on mental health, builds a vector store index of embeddings, and retrieves answers with a query engine.

Format Prompt Templates19:49
Explore dynamic formatting of prompt templates in a RAC application with Lamar Index library. Master partial formatting, prompt variable mapping, and function-based formatting to tailor prompts with context and tone.
Conversational Prompts11:43
Semantic Similarity Evaluator8:51
Demonstrates a semantic similarity evaluator by converting two texts to language embeddings and measuring distance to produce a similarity score. Shows that a score above 0.8 indicates similarity.
Language Embeddings and Vector Databases40:34
Using a Chroma DB Vector Database19:51
Learn to use chroma db as a persisted vector store for simple directory reader-based rag workflows, compare openai and gemini models, and control context window and output size.
LlamaIndex with SQL Database24:06
Explore three ways to query a MySQL Walmart dataset with llama index: a SQL query engine, a SQL table retrieval engine, and a natural language SQL retriever.
LlamaIndex Query Pipelines15:47
LlamaIndex query pipelines organize prompts, language models, retrieval, rerankers, and response synthesizers into modular sequential or dag workflows for rag applications, enabling observability and evaluation.
Setting up a Simple Sequential Query Pipeline6:38
Learn to build a simple sequential query pipeline with LlamaIndex, using a prompt and GPT-3.5-turbo to retrieve a movie's name, cast, and year.
Setting up a DAG Pipeline21:27
Explore how to build a complex dag pipeline for gen ai rag using LlamaIndex, integrating a retriever, Cohere reranker, and tree summarize to derive concise insights from multiple documents.
Setting up a Dataframe Pipeline22:40
Build a dataframe pipeline in LlamaIndex to run pandas queries on an obesity dataset, using a data frame source and a LM workflow to compute correlations and synthesize results.
Working with Agents and Tools14:05
Explore how llama index agents and tools enable intelligent decisions for external data sources. Define descriptive tool names, orchestrate tool use, and leverage Lama Hub for pre-built tools.
Create a Calculator using a ReAct Agent12:37
Create a Document Agent with Dynamically built Tools21:51
Build an adaptive document agent with dynamically created tools to query three pdf documents via vector indexes and retrievers, enabling tool selection and multi-source answer synthesis.
Build a Code Checker with Streamlit UI14:36
Build a Streamlit UI for a RAG app and code checker powered by LlamaIndex and Gemini LM, letting users select a language, enter a task and code, and validate it.

Requirements

Some Python background

Description

This course uses Open AI GPT and Google Gemini APIs, LlamaIndex LLM Framework and Vector Databases like ChromaDB and Pinecone, and is intended to help you learn how to build LLM RAG applications through solid conceptual and hands-on sessions. This course covers all the basic aspects to learn LLM RAG apps and Frameworks like Agents, Tools, QueryPipelines, Retrievers, Query Engines in a crisp and clear manner. It also takes a dive into concepts of Language Embeddings and Vector Databases to help you develop efficient semantic search and semantic similarity based RAG Applications. We will also cover multiple Prompt Engineering techniques that will help make your RAG Applications more efficient.

List of Projects/Hands-on included:

Basic RAG: Chat with multiple PDF documents using VectorStore, Retriever, Nodepostprocessor, ResponseSynthesizer and Query Engine.

ReAct Agent: Create a Calculator using a ReAct Agent and Tools.

Document Agent with Dynamic Tools : Create multiple QueryEngineTools dynamically and Orchestrate queries through Agent.

Semantic Similarity: Try Semantic Similarity operations and get Similarity Score.

Sequential Query Pipeline: Create Simple Sequential Query Pipeline.

DAG Pipeline: Develop complex DAG Pipelines.

Dataframe Pipeline: Develop complex Dataframe Analysis Pipelines with Pandas Output Parser and Response Synthesizer.

Working with SQL Databases: Develop SQL Database ingestion bots using multiple approaches.

For each project, you will learn:

- The Business Problem

- What LLM and LlamaIndex Components are used

- Analyze outcomes

- What are other similar use cases you can solve with a similar approach.

Who this course is for:

Software Developers aspiring to use the power of LLMs to build Gan AI RAG Applications as part of their Project and Products
Software Developers looking to automate their Software Engineering processes using Gen AI

Gen AI - RAG Application Development using LlamaIndex

What you'll learn

Explore related topics

Course content

Introduction7 lectures • 3hr 23min

Getting Deeper into LlamaIndex14 lectures • 4hr 15min

Requirements

Description

Who this course is for: