Teach on Udemy

Turn what you know into an opportunity and reach millions around the world.

Learn More

Your cart is empty.

Keep shopping

AI Document Intelligence: RAG, Agents & ML Data

Name: AI Document Intelligence: RAG, Agents & ML Data
Rating: 4.8 (29 reviews)

Build a complete AI app that turns PDFs into RAG answers, agentic insights, APIs, UI, and ML-ready data.

Highest Rated

New

Created byRahul Sahay

Last updated 6/2026

English

What you'll learn

Build an end-to-end AI Document Intelligence platform using PDFs, RAG, AI Agents, FastAPI, React, and ML-ready datasets.
Implement Retrieval-Augmented Generation (RAG) using ChromaDB, embeddings, semantic search, and grounded AI responses.
Design and develop AI Agents that select tools, execute queries, and generate intelligent responses from documents and data.
Transform unstructured PDF documents into structured, analytics-ready, and machine-learning-ready datasets.
Build production-style FastAPI backend services and expose AI, RAG, and Agent capabilities through REST APIs.
Develop a modern React application to interact with RAG pipelines, AI Agents, and document intelligence workflows.
Implement document ingestion, text extraction, cleaning, chunking, embeddings, and vector database pipelines from scratch.
Apply real-world AI architecture patterns used in enterprise document intelligence, healthcare, and knowledge systems.

Course content

9 sections • 91 lectures • 11h 53m total length

Introduction5:41
About the Project9:02
Explore the architecture of a production-ready ai document intelligence project that turns unstructured claim pdfs into structured data using rag, embeddings, and a guarded, auditable agentic flow.
Understanding Folder Structure8:15
Understanding Readme File9:13
Github Strategy4:13
About Me4:50
Meet Rahul Sahay, a transformational leader and enterprise AI architect who leads a Bupa program from replatforming to modernization and shares expertise in data strategy and machine learning.

Introduction2:21
Explore the document intelligence pipeline by loading chunks as embeddings into a chroma vector store, then search via cosine similarity to retrieve the most relevant chunks for queries.
Creating Chunk Loader Helper Methods9:07
Load chunk files and validate each chunk to ensure required fields, then build a chunk manifest for rag input. Save validated chunks to a vector db for rag search.
Loading Valid Chunks6:17
Load all chunks from the chunks directory, locate and validate chunk records from chunk json files, and return the collection of valid chunks while preparing the manifest.
Creating the Chunk Manifest File6:47
Saving the Chunk Manifest File4:45
Wiring up with Main File5:28
Fixing the Manifest Creation Issue4:02
Creating AppSettings File7:42
How Cosine Similarity Powers AI Search6:03
Creating Embedding Client and Chroma Collection8:41
Creating helper methods to check chunk existence10:40
Creating Chroma Db Vector Store10:03
Searching the Vector Store12:23
Leverage vector store search to implement a robust document retrieval pipeline using top-k queries, embedding clients, and chroma collections, returning documents, metadata, and distances for sanity testing.
Printing the search results3:45
Creating API Keys and Wiring up Main File9:24
Vector Store Demo5:34

Introduction4:54
Apply a retrieval augmented generation workflow by embedding chunks in a vector store, use an LLM to generate grounded answers, and cache results in a json file to save tokens.
Creating Caching Layer for QA Pipeline11:34
Build a qa pipeline with a caching layer that normalizes questions for cache keys, loads cached answers, and uses vector store search when cache misses.
Building Source References using Metadata9:33
Building the Relevant Context9:49
Building the RAG Prompt6:12
Building RAG Q&A with Context8:35
Wiring up RAG Q&A with and without context8:30
Printing RAG Answer7:03
learn to print and validate a rag answer by handling rag responses, extracting questions, answers, and sources, and displaying chunk details and cosine distances for debugging.
Saving the RAG Answers to Cache5:06
Wiring up the Main File8:11
Fixing the Main File Issue2:36
RAG Q&A Demo4:59
Ground the rag Q&A in a single claim document, extracting the total claim amount, diagnosis, and policy number while filtering by claim ID to prevent hallucination.

Introduction2:20
Creating the Claim Schema using Pydantic7:05
Creating method for getting and saving schema5:29
Implement get claim schema to return the json schema for the claim record model and save claim extraction as a json file to an output path using json.dumps with utf-8.
Wiring up Main file with Claim Schema3:48
Building Extraction Prompt and Loading Claim Records13:03
Build a claim extractor that loads cached claim records and validates them against the claim schema. Create a structured extraction prompt that outputs JSON with nulls for missing fields.
Extracting Claim Records from Text8:21
Extracting Claim Records from Cleaned Claims11:30
Extract claim records from text and convert them to JSON for validation, then perform batch processing over cleaned claims with duplicate checks and structured record creation.
Saving the Records as JSON and CSV11:09
Wiring up Main File to generate ML Ready CSV Dataset8:21
Wire up the main file to extract and save claim records as an ml-ready csv dataset, exporting validated structured claim records for ml pipelines.
Validating Claim Record12:59
Validating Claim Records10:16
Saving the Data Quality Report3:34
Wiring up Data Quality Report7:20
Creating Helper Methods for Data Dictionary6:44
Creating Data Dictionary9:54
Wiring Up Data Dictionary with Main file5:59
Building Processing Summary9:23
Wiring up Processing Summary with Main9:13

Introduction2:37
Learn how rule-based and llm-based agents route requests, select tools like summarize pipeline outputs and query claim data set, and perform rag-based document lookups for fast, token-efficient request handling.
Creating Document Tools Helper Methods15:47
Explore ai document intelligence by building agentic tools that orchestrate rag pipelines, vector db retrieval, and claim processing with helper methods for data normalization and json loading.
Creating Document Methods for AI Agent14:17
Creating Query Operators and Filters18:03
Creating Query Claim Dataset & Ask Documents5:26
Creating a Simple AI Agent11:17
Build a simple ai agent that uses intent detection to route requests to document tools for data quality, pipeline summaries, and status filtering in an ai document intelligence workflow.
Print & Saving Agent Responses6:47
Describe how to print and save agent responses by detecting intent, printing the response and the tool used, and saving the results to an agent responses.json file with utf-8 encoding.
Building LLM AI Agent7:54
Running a Rule Based AI Agent Flow9:44
Fixing Rule Based AI Agent3:42
Running LLM Agent5:20

Requirements

Basic Python programming knowledge is recommended, but every concept is explained step-by-step.
No prior experience with RAG, AI Agents, ChromaDB, FastAPI, React, or Document Intelligence is required.
A computer capable of running Python applications and installing open-source packages.
Basic understanding of APIs, JSON, and software development concepts will be helpful.
A willingness to build real-world AI applications through hands-on project-based learning.
No prior Machine Learning or Data Science experience is required.
No prior experience with Vector Databases or LLM frameworks is needed.
Students should be comfortable using VS Code or any Python development environment.

Description

AI Document Intelligence: RAG, Agents & ML Data

Build a complete AI-powered Document Intelligence platform from scratch and learn how to transform unstructured PDFs into intelligent applications, structured datasets, AI agents, and ML-ready data.

Most AI courses stop at embeddings and question-answering. This course goes much further.

You will build an end-to-end healthcare claims intelligence platform that starts with raw PDF documents and evolves into a production-style system featuring RAG, AI Agents, FastAPI services, React applications, structured datasets, analytics-ready outputs, and machine learning pipelines.

Throughout the course, you will work on a realistic project and implement every major component yourself instead of relying on black-box frameworks.

What You Will Build

PDF ingestion and document processing pipeline
Automated text extraction from real-world documents
Data cleaning and preprocessing workflows
Intelligent document chunking strategies
Embedding generation and vector storage using ChromaDB
Retrieval-Augmented Generation (RAG) applications
AI Agents capable of selecting and executing tools
Structured claim datasets generated from unstructured documents
ML-ready datasets for analytics and machine learning
FastAPI backend services
Modern React frontend application
End-to-end AI Document Intelligence platform

What You Will Learn

Document Intelligence architecture and design patterns
RAG implementation from scratch
Vector databases and semantic search
ChromaDB integration
Prompt engineering for retrieval systems
Agentic AI workflows and tool usage
Dynamic query planning and execution
Structured data extraction from PDFs
Data quality validation and reporting
FastAPI API development
React application development
Building production-style AI applications
Preparing data for Machine Learning and MLOps workflows

Why This Course Is Different

Most courses teach RAG as an isolated concept.

This course demonstrates how RAG fits into a complete AI ecosystem where documents are processed, validated, transformed into structured data, queried through AI agents, exposed through APIs, visualized in modern web applications, and ultimately prepared for machine learning use cases.

You will understand not only how individual components work, but also how they fit together to create enterprise-grade AI solutions.

Course Statistics

11.5+ Hours of Content
91+ Lectures
End-to-End Project-Based Learning
Real-World Healthcare Claims Use Case
FastAPI + React Integration
RAG + Agents + ML Data Pipeline
Source Code Included

Who This Course Is For

AI Engineers
Machine Learning Engineers
Data Scientists
Python Developers
Full Stack Developers
Solution Architects
GenAI Practitioners
Students looking to build real-world AI applications

Prerequisites

Basic Python knowledge
Basic understanding of APIs
Curiosity to learn AI, RAG, Agents, and Document Intelligence

By the end of this course, you will have built a complete AI Document Intelligence platform capable of transforming raw PDFs into searchable knowledge, intelligent agent workflows, structured datasets, analytics-ready outputs, and ML-ready data pipelines.

Who this course is for:

Python developers who want to build real-world RAG and Agentic AI applications.
AI Engineers and GenAI practitioners looking to move beyond basic chatbot implementations.
Machine Learning Engineers who want to create ML-ready datasets from unstructured documents.
Data Scientists interested in Document Intelligence, knowledge extraction, and AI automation.
Full Stack Developers who want to integrate FastAPI, React, RAG, and AI Agents into modern applications.
Software Architects and Technical Leads exploring enterprise AI solution design patterns.
Backend Developers interested in vector databases, semantic search, and AI-powered APIs.
Anyone looking to build end-to-end AI Document Intelligence platforms using RAG, Agents, FastAPI, React, and structured data pipelines.

AI Document Intelligence: RAG, Agents & ML Data

What you'll learn

Explore related topics

Course content

Introduction6 lectures • 41min

Project Setup and Document Processing11 lectures • 1hr 32min

Preparing For RAG7 lectures • 1hr 1min

Working with Vector Store & Embeddings16 lectures • 1hr 53min

Building RAG Question & Answer Pipeline12 lectures • 1hr 27min

Making Data ML Ready18 lectures • 2hr 26min

Building AI Agents11 lectures • 1hr 41min

Building Fast API Layer and Connecting with React App9 lectures • 1hr 9min

Bonus Section1 lecture • 3min

Requirements

Description

Who this course is for: