---

🏠 Home - All Lessons

⬅️ Previous - Welcome to Agentic AI Essentials
➡️ Next - Program FAQs

---

The Journey Ahead

This program focuses on the foundations of agentic AI and retrieval-augmented systems.

You’ll learn how agentic AI systems are structured, how to control LLM behavior using prompts and memory, and how to ground AI responses in your own data using retrieval and vector search. The program culminates in building a complete RAG-powered AI assistant, designed to be understandable, inspectable, and easy to extend.

---

Program Structure

The program is organized into four units, each building a specific capability required for agentic AI development.

• Unit 1 establishes the mental models behind agentic systems
• Unit 2 focuses on controlling LLM behavior through prompts, reasoning, and memory
• Unit 3 teaches how to ground AI responses in custom data using retrieval and vector search
• Unit 4 is a capstone project where everything comes together

Each unit prepares you for the next. Nothing is included out of sequence.

---

Tools & Frameworks You’ll Work With

You’ll use modern, widely adopted tools — with an emphasis on concepts that transfer beyond any single library.

LLMs & Providers

• Groq (primary, free-tier friendly)
• Ollama for running local LLMs
• OpenAI and other providers (discussed as alternatives)

Frameworks & Libraries

• LangChain for prompting, chaining, memory, and RAG
• Pydantic for structured output validation

Retrieval & Storage

• ChromaDB (primary vector database)
• FAISS and other options (discussed conceptually)

Embedding Models

• Hugging Face sentence-transformer models
• OpenAI embeddings (discussed as alternatives)

Core Technologies

• Python
• Embedding models for semantic search
• Retrieval-Augmented Generation (RAG) patterns

Where relevant, you’ll also learn why certain tools are chosen and what tradeoffs they involve.

---

What You’ll Build

This program culminates in a capstone project:

You’ll build a RAG-powered AI assistant that:

• retrieves information from a custom document set,
• uses structured prompts and memory,
• and generates grounded, reliable responses.

The capstone is supported by:

• a clear set of requirements,
• a starter code template,
• and explicit submission guidance.

You’re never guessing what “complete” means.

---

Final Note

This curriculum is designed to help you build real capability, not just familiarity with tools.

If you want a high-level overview, scan the curriculum map below.
If you want to understand how the pieces connect, this page is your guide.

When you’re ready, continue to Unit 1: Understanding Agentic AI and begin the journey.

---

🏠 Home - All Lessons

⬅️ Previous - Welcome to Agentic AI Essentials
➡️ Next - Program FAQs

---