Ready Tensor Agentic AI Certification - Week 2

This repository contains the lessons, practical code and exercises for Week 2 of the Agentic AI Developer Certification Program by Ready Tensor, covering foundational prompt engineering concepts that are essential for building effective agentic AI systems.

What You'll Learn

• How to build modular prompts for better clarity and reuse
• When and how to apply reasoning techniques like CoT, ReAct, and Self-Ask
• Strategies for structured output parsing
• Principles of function chaining in AI workflows
• How vector databases enable semantic search
• Foundations of Retrieval-Augmented Generation (RAG)

---

Lessons in This Repository

0. Getting Started: Free APIs and Local LLMs

Set up your environment with free LLM options — including cloud APIs like Groq and Google Gemini, or local models via Ollama — so you can follow the course without hitting cost barriers.

1a. Building Prompts for Agentic AI Systems

Learn how to design effective prompts using modular components — instruction, tone, role, constraints — and how to iteratively refine prompts for clarity and consistency.

1b. Prompt Engineering: Advanced Reasoning Techniques

Covers three powerful techniques — Chain of Thought, ReAct, and Self-Ask — and shows how to incorporate them into your modular prompt framework.

2. From Text to Data: Hands-On LLM Output Parsing

Explore structured output generation from LLMs using prompt formatting and model-native methods, with tools like Pydantic and LangChain.

3. Function Chaining for Intelligent Pipelines

Understand how breaking down tasks into smaller functions enables composable, robust AI systems, and how to structure chains for clarity and reliability.

4a. Vector Databases: Finding Meaning, Not Just Keywords

Get introduced to vector search and how embeddings power semantic retrieval in intelligent systems.

4b. Vector Databases: Building a Semantic Retrieval System

Build a working pipeline with ChromaDB, embeddings, and chunked documents — the foundation of modern RAG workflows.

5. Introduction to RAG (Retrieval Augmented Generation)

Learn why RAG outperforms fine-tuning for most real-world scenarios, and how it enables domain-specific, knowledge-grounded assistants.

Repository Structure

rt-agentic-ai-cert-week2/
├── code/
│   ├── config/
│   │   ├── config.yaml          # App config with reasoning strategies
│   │   └── prompt_config.yaml   # Prompt configurations for examples
│   ├── lesson_1a_and_ab.py      # Main script for lesson 1
│   ├── lesson_2.py      # Main script for lesson 2
│   ├── paths.py                 # File path configurations
│   ├── prompt_builder.py        # Modular prompt construction functions
│   └── utils.py                 # Utility functions
├── data/
│   └── vae-publication.md       # Sample publication for exercises
├── lessons/
│   └── lesson-wk2-*             # Markdown files and visuals for each lesson
├── outputs/                     # Generated prompts and LLM responses
├── .env.template                # Environment variables template
├── requirements.txt             # Python dependencies
└── README.md

Installation & Setup

1. Clone the repository:

   git clone https://github.com/readytensor/rt-agentic-ai-cert-week2.git
   cd rt-agentic-ai-cert-week2

2. Install dependencies:

   pip install -r requirements.txt

3. Set up your API key:

   Create a .env file in the root directory and add at least one API key. You need at least one key from OpenAI, Groq, or Google to run the examples.

   See .env.example file for the complete template.

   # Choose at least one (you don't need all three)
   OPENAI_API_KEY=your-openai-key-here
   GROQ_API_KEY=your-groq-key-here
   GOOGLE_API_KEY=your-google-key-here

   Get your free API key from:

   • OpenAI (paid): platform.openai.com/api-keys
   • Groq (free): console.groq.com
   • Google Gemini (free): makersuite.google.com

   💡 No budget for APIs? Check out our Free API Setup Guide for using Groq and Google's free tiers!

4. Run the examples:

   cd code
   python lesson_1_and_2.py

   Customize your experiments: Edit the prompt_cfg_key variable in lesson_1_and_2.py (near the bottom of the script) to test different prompt configurations (e.g., summarization_prompt_cfg1 through summarization_prompt_cfg6). You can also create new configurations in config/prompt_config.yaml to experiment with your own prompt designs.

Key Features

• 🧩 Modular Prompt Builder Construct prompts from reusable components like role, constraints, tone, and goals.

• 🧠 Reasoning Techniques Integrate CoT, ReAct, and Self-Ask patterns into your prompts with a single config line.

• 🛠️ Hands-On Output Parsing Generate and validate structured output using both prompt-based and model-native methods.

• 🔗 Function Chaining Framework Chain modular steps to build robust, maintainable AI pipelines.

• 🔍 Vector Search Implementation Build a retrieval system using real embeddings and ChromaDB.

• 🔍 Intro to RAG Introduction to Retrieval-Augmented Generation.

License

This project is licensed under the CC BY-NC-SA 4.0 License - see the LICENSE file for details.

Contact

Ready Tensor, Inc.

• Email: contact at readytensor dot com
• Issues & Contributions: Open an issue or pull request on this repository
• Website: Ready Tensor