33 MODULES • 9 TRACKS • 20 CAPSTONE LABS

Building AI Agents with Claude

From Hello World to Autonomous Production Systems. Master the full agent lifecycle — design, build, protect, observe, and deploy — with hands-on code in Python and TypeScript.

🔍 Search the Course  ·  💬 Ask the Course (works offline)

33
Modules
22
Capstones
33 of 33 modules complete (100%)

On a phone or tablet? Try the mobile-optimized versions with condensed content and pseudocode.

View Mobile Versions

Track 1 — Foundations

7 modules
M00 The Agent Lifecycle — See the Whole Picture First
Beginner45 min
 M00B Hello World — Three Ways to Build an Agent
Hands-On30-45 min
 M01 The LLM Mental Model
Beginner60 min
 M02 Tokens — The Atoms of AI Communication
Beginner50 min
 M03 Prompts — Programming in Natural Language
Beginner60 min
 M03B Context Engineering — Curating What the Model Sees
Beginner→Intermediate50 min
 M04 Structured Output & Parsing
Beginner55 min

Track 2 — Tool Use

3 modules
M05 Function Calling Fundamentals
Intermediate65 min
 M06 Multi-Tool Orchestration
Intermediate70 min
 M07 MCP — Model Context Protocol
Intermediate70 min

Track 3 — Memory & RAG

4 modules
M08 Conversation Management
Intermediate60 min
 M09 RAG — Retrieval-Augmented Generation
Intermediate75 min
 M10 Advanced RAG Patterns
Advanced80 min
 M11 Multi-Layer Memory Architecture
Advanced75 min

Track 4 — Agent Architecture

5 modules
M12 The ReAct Pattern — Reason, Act, Observe
Intermediate70 min
 M13 Planning & Task Decomposition
Advanced75 min
 M14 Multi-Agent Systems
Advanced80 min
 M15 Code Interpreter & Sandbox Execution
Advanced70 min
 M15B BUILD: Complete Agent & Subagent System
Hands-On Lab120 min

Track 5 — Guardrails & Safety

3 modules
M16 Input Guardrails
Intermediate65 min
 M17 Output Guardrails & Human-in-the-Loop
Intermediate70 min
 M18 Evaluation & Testing
Advanced75 min

Track 6 — Observability

2 modules
M19 Tracing & Logging
Intermediate65 min
 M20 Monitoring & Continuous Improvement
Advanced70 min

Track 7 — Deployment & Operations

6 modules
M21 API Design & Deployment
Advanced75 min
 M21B Cloud Provider Deep Dive — Bedrock, Vertex AI & Azure AI
Advanced90 min
 M22 Cost Optimization
Advanced70 min
 M22B BUILD: Deploy Your Agent — Local, GCP & AWS
Hands-On Lab120 min
 M21C Headless Agents — Running Without a Human
Advanced55 min
 M22C Agentic Coding Quality Gates & Rich Output
Advanced60 min

Track 8 — Capstones

2 modules + 22 capstone labs
M23 Capstone Project Series — Overview & Guide
Advanced30 min
 M24 What's Next — The Agent Frontier
All Levels45 min
Domains: A — Healthcare Pre-Auth B — B2B Ecommerce C — Public Records / UCC
Capstone 1 — Your First Agent (after Track 2)
CAPSTONE-1 / DOMAIN-A Healthcare

Pre-Auth Status Checker

 CAPSTONE-1 / DOMAIN-B B2B Ecommerce

Order Status Bot

 CAPSTONE-1 / DOMAIN-C Public Records

UCC Filing Lookup Agent
Capstone 2 — RAG-Powered Q&A (after Track 3)
CAPSTONE-2 / DOMAIN-A Healthcare

Clinical Policy Q&A System

 CAPSTONE-2 / DOMAIN-B B2B Ecommerce

Product Catalog & Contract Q&A

 CAPSTONE-2 / DOMAIN-C Public Records

UCC Regulatory Reference Agent
Capstone 3 — Autonomous Agent (after Track 4)
CAPSTONE-3 / DOMAIN-A Healthcare

Pre-Auth Decision Support Agent

 CAPSTONE-3 / DOMAIN-B B2B Ecommerce

Order Exception Resolution Agent

 CAPSTONE-3 / DOMAIN-C Public Records

Entity Resolution Agent
Capstone 4 — End-to-End Pipeline (after Track 6)
CAPSTONE-4 / DOMAIN-A Healthcare

End-to-End Pre-Auth Pipeline

 CAPSTONE-4 / DOMAIN-B B2B Ecommerce

B2B Order Lifecycle Pipeline

 CAPSTONE-4 / DOMAIN-C Public Records

UCC Data Pipeline Orchestration
Capstone 5 — Production System (after Track 7)
CAPSTONE-5 / DOMAIN-A Healthcare

Production Pre-Auth System

 CAPSTONE-5 / DOMAIN-B B2B Ecommerce

Production B2B Order System

 CAPSTONE-5 / DOMAIN-C Public Records

Production UCC Data Intelligence Platform
Capstone 6 — Bonus Lab
CAPSTONE-6 / BONUS Data Engineering

Parallel State Testing Agent — Bronze Canonical Load Validation
Capstone 7 — Agent Evolution: Same Agent, Three Ways (graduation project)
CAPSTONE-7 / DOMAIN-A Healthcare

Pre-Auth Decision Agent — Raw → SDK → Spec-Driven

 CAPSTONE-7 / DOMAIN-B B2B Ecommerce

Order Exception Agent — Raw → SDK → Spec-Driven

 CAPSTONE-7 / DOMAIN-C Public Records

UCC Filing Risk Analyzer — Raw → SDK → Spec-Driven
Capstone 8 — Bonus Lab (legacy modernization)
CAPSTONE-8 / BONUS Data Engineering

Legacy Migration Agent — Oracle → PostgreSQL

 CAPSTONE-8B / BONUS Agent Skills

Legacy Migration Agent — Skills-First

Track 9 — Certification Prep

4 modules
M25 Claude Code Mastery
Advanced75 min
 M26 Hooks, Sessions & the Agent SDK
Advanced80 min
 M27 Certification Exam Prep
Advanced90 min
 M27B Cert Domain 5.5+5.6 Deep Dive
Intermediate→Advanced65 min

Supplementary Material

7 pages

The appendices fill in language and framework background that the modules assume you already have — Appendix A is the de facto prerequisite for M01, so start there if venv, pip, or JSON are unfamiliar. The bonus pages are genuinely optional and go wider than the course needs. If you have never used an AI coding tool on a real codebase, take Code With AI before this course.

APPENDIX A Python Essentials for This Course
Beginner45 min
 APPENDIX B FastAPI Essentials for Agent Deployment
Beginner40 min
 APPENDIX C PySpark Essentials for Data Pipeline Agents
Intermediate50 min
 APPENDIX D Claude Code with Open Source & Local Models
Intermediate40 min
 BONUS Generative AI Foundations
Beginner45 min
 BONUS Building Effective Agents
Intermediate60 min
 BONUS Jev — The Cheap Judge in a Claude Code Loop
Intermediate70 min + lab
 M28 Open Source Models — Agents Without an API Key
Beginner→Intermediate75 min

Choose Your Learning Path

Three curated paths through the curriculum based on your goals and available time.

Path A — Weekend Builder

Fastest path to a working agent
Best for: Developers who want to ship something this weekend
~12 modules • ~13 hours

Path B — Deep Diver

Comprehensive understanding of every concept
Best for: Learners who want to deeply understand agent internals
~30 modules + 6 capstones • ~50 hours

Path C — Production Engineer

Reliability, observability, and deployment focus
Best for: Engineers shipping agents to production with SLAs
~16 modules + 4 capstones • ~35 hours

Preparing for the Claude Certified Architect exam? Track 9 (M25–M27B) layers on top of any path.

Capstone Projects

5 progressive capstones across 3 industry domains, plus 1 bonus lab and a 3-domain graduation project (Capstone 7). Pick one domain and follow it through all five levels, or mix and match.

Domains: A — Healthcare Pre-Authorization (CPT/ICD codes, HIPAA, clinical criteria) B — B2B Ecommerce (PO lifecycle, carrier APIs, SLA management) C — Public Records (UCC liens, entity resolution, Medallion Architecture)

Capstone 1 — Your First Agent (after Track 2)

CAPSTONE-1 / DOMAIN-A Healthcare

Pre-Auth Status Checker

 CAPSTONE-1 / DOMAIN-B B2B Ecommerce

Order Status Bot

 CAPSTONE-1 / DOMAIN-C Public Records

UCC Filing Lookup Agent

Capstone 2 — RAG-Powered Q&A (after Track 3)

CAPSTONE-2 / DOMAIN-A Healthcare

Clinical Policy Q&A System

 CAPSTONE-2 / DOMAIN-B B2B Ecommerce

Product Catalog & Contract Q&A

 CAPSTONE-2 / DOMAIN-C Public Records

UCC Regulatory Reference Agent

Capstone 3 — Autonomous Agent (after Track 4)

CAPSTONE-3 / DOMAIN-A Healthcare

Pre-Auth Decision Support Agent

 CAPSTONE-3 / DOMAIN-B B2B Ecommerce

Order Exception Resolution Agent

 CAPSTONE-3 / DOMAIN-C Public Records

Entity Resolution Agent

Capstone 4 — End-to-End Pipeline (after Track 6)

CAPSTONE-4 / DOMAIN-A Healthcare

End-to-End Pre-Auth Pipeline

 CAPSTONE-4 / DOMAIN-B B2B Ecommerce

B2B Order Lifecycle Pipeline

 CAPSTONE-4 / DOMAIN-C Public Records

UCC Data Pipeline Orchestration

Capstone 5 — Production System (after Track 7)

CAPSTONE-5 / DOMAIN-A Healthcare

Production Pre-Auth System

 CAPSTONE-5 / DOMAIN-B B2B Ecommerce

Production B2B Order System

 CAPSTONE-5 / DOMAIN-C Public Records

Production UCC Data Intelligence Platform

Capstone 6 — Bonus: Data Pipeline Testing

CAPSTONE-6 / BONUS Data Engineering

Parallel State Testing Agent — Bronze Canonical Load Validation

Capstone 7 — Agent Evolution: Same Agent, Three Ways (graduation project)

Build the same agent three times: once with the raw API loop, once with claude-agent-sdk, and once spec-driven via Claude Code. Compare code size, dev time, flexibility, and control across all three. The capstone version of M00B.

CAPSTONE-7 / DOMAIN-A Healthcare

Pre-Auth Decision Agent — Raw → SDK → Spec-Driven

 CAPSTONE-7 / DOMAIN-B B2B Ecommerce

Order Exception Agent — Raw → SDK → Spec-Driven

 CAPSTONE-7 / DOMAIN-C Public Records

UCC Filing Risk Analyzer — Raw → SDK → Spec-Driven

Capstone 8 — Legacy Modernization (bonus lab)

Migrate a legacy Oracle database to PostgreSQL 16 — schema, data, PL/SQL and application SQL — against two live containers. A coordinator plus five specialists, a read-only source enforced in code, a validator that proves its own work, and a cutover gate only a human can open. Then Capstone 8B rebuilds the identical problem with Agent Skills instead of subagents, and measures which architecture actually did better.

CAPSTONE-8 / BONUS Data Engineering

Legacy Migration Agent — Oracle → PostgreSQL

 CAPSTONE-8B / BONUS Agent Skills

Legacy Migration Agent — Skills-First

Prerequisites Map

Module dependencies — complete prerequisites before moving on.

ModuleRequiresUnlocks
M00 — Course OverviewNoneM00B, M01
M00B — Hello World (Three Approaches)M00M01 (concepts), M15B + Capstone 7 (full pattern)
M01 — LLM Mental ModelM00M02, M03
M02 — TokensM01M22 (Cost)
M03 — PromptsM01M03B, M04, M05
M03B — Context EngineeringM03M04, M08
M04 — Structured OutputM03M05
M05 — Function CallingM03, M04M06, Capstone 1
M06 — Multi-Tool OrchestrationM05M07, M12
M07 — MCPM06M08
M08 — Conversation ManagementM07M09
M09 — RAG FundamentalsM08M10, Capstone 2
M10 — Advanced RAGM09M11
M11 — Multi-Layer MemoryM10M12
M12 — ReAct Agent LoopM06, M11M13
M13 — Planning & DecompositionM12M14
M14 — Multi-Agent SystemsM13M15, Capstone 3
M15 — Code InterpreterM14M15B
M15B — BUILD: Agent SystemM15M16
M16 — Input GuardrailsM15BM17
M17 — Output Guardrails & HITLM16M18
M18 — Evaluation & TestingM17M19, Capstone 4
M19 — Tracing & LoggingM18M20
M20 — Monitoring & CIM19M21
M21 — API Design & DeploymentM20M22
M21B — Cloud Provider Deep DiveM21M22
M22 — Cost OptimizationM02, M21M22B
M22B — BUILD: Deploy to CloudM22M22C, Capstone 5
M21C — Headless AgentsM22BM22C
M22C — Agentic Coding Quality GatesM22BM23
M23 — Capstone Series GuideM22CAll capstones
M24 — The Agent FrontierNone (standalone)M25
M25 — Claude Code MasteryM24M26
M26 — Hooks, Sessions & SDKM25M27
M27 — Cert Exam PrepM26M27B
M27B — Cert Domain 5.5+5.6 Deep DiveM09, M11, M17, M18, M27Exam ready!
Capstone 7 — Agent Evolution (graduation)M15B, M25, M26 (M01–M27 recommended)Course complete
Capstone 8 — Legacy Migration (bonus)M07, M13–M14, M15B, M16–M18, M22BCourse complete
Capstone 8B — Skills-First (bonus)Capstone 8, M25–M26Course complete

📂 The labs and sample projects for this course live in github.com/varasrinivas/agenticai-course — git clone https://github.com/varasrinivas/agenticai-course.git