Generative AI with Python

Packt via Coursera

Go to class Write review

Overview

Coursera Flash Sale

40% Off Coursera Plus for 3 Months!

Grab it

This course features Coursera Coach! A smarter way to learn with interactive, real-time conversations that help you test your knowledge, challenge assumptions, and deepen your understanding as you progress through the course. Unlock the power of generative AI by mastering Python and working hands-on with cutting-edge tools and libraries. From building large language models (LLMs) to implementing advanced agentic systems, this course takes you on an in-depth journey through AI development. You’ll explore the essentials of LLMs, model training, parameter tuning, and the integration of advanced techniques like Retrieval-Augmented Generation (RAG) and vector databases. The interactive learning experience ensures you are not just passively absorbing information but engaging with practical coding exercises and real-world applications. The course begins with the foundational setup, including Python, IDEs, and environment configurations, before diving deep into LLMs, multimodal models, and even exploring agent-based systems. You’ll move through advanced topics such as prompt crafting, chaining models, and building intelligent systems with frameworks like crewAI and AG2. The journey concludes with model fine-tuning techniques, including Low-Rank Adaptation (LoRA), that enable you to optimize performance. This course is designed for AI enthusiasts, data scientists, and developers who want to expand their skills in generative AI. It is ideal for anyone with basic knowledge of Python who wants to build AI-driven applications. The course is suitable for those at an Intermediate level with some prior programming experience in Python. By the end of the course, you will be able to design and implement generative AI models, create complex AI workflows using chains and agents, manage vector databases, and fine-tune models to suit specific tasks and domains.

Syllabus

Course Introduction

In this module, we will introduce the course and provide an overview of the instructor’s background in AI and Python. We will explore the course objectives and structure to ensure you know what to expect. Additionally, we’ll guide you through the essential system setup, including installing tools like Python, an IDE, and managing API keys for the hands-on coding exercises.

Large Language Models – Introduction

In this module, we will explore the foundational concepts of Large Language Models (LLMs) and how they function within the AI space. We will compare traditional NLP techniques with LLMs to understand their advancements. Additionally, we will evaluate the real-world achievements and performance of these models across different tasks.

Large Language Models – Deep Dive

In this module, we will dive deep into the training process of Large Language Models, uncovering the complexities of data preparation and optimization techniques. We will explore ways to improve model performance and evaluate major LLM providers and their products. Additionally, you will learn how to interact with different LLMs via hands-on coding exercises.

Large Language Models – Types and Variants

In this module, we will explore various types of Large Language Models, including how to run models locally on your system. You will also dive into multimodal models, which combine text, images, and other media to enhance AI capabilities. Additionally, we will look at tokenization methods and how they support AI systems in processing and understanding data inputs.

Large Language Models – Chains

In this module, we will introduce you to the concept of chains in AI, where multiple model interactions are linked together to form complex workflows. You will learn how to design and implement prompt templates for repeated use cases, and create systems where outputs are structured and can adapt based on different decision branches in your application.

Vector Databases

In this module, we will explore vector databases and their significance in managing and retrieving high-dimensional data for AI applications. You will learn to work with vector embeddings, chunk data for more efficient storage, and practice querying databases to retrieve relevant information based on similarity searches.

Retrieval-Augmented Generation – Baseline

In this module, we will introduce you to Retrieval-Augmented Generation (RAG) and walk you through its core phases, from data retrieval to response generation. You will gain hands-on experience in coding a basic RAG pipeline, enhancing the accuracy and relevance of the AI outputs by incorporating external information into the model’s process.

Retrieval-Augmented Generation – Advanced

In this module, we will take a deeper dive into advanced techniques for enhancing Retrieval-Augmented Generation workflows. You will learn how to optimize data retrieval and refine responses with strategies like query expansion, prompt compression, and speculative RAG. Additionally, we will explore multimodal RAG and hybrid approaches to handle diverse data types efficiently.

Agentic Systems – Overview

In this module, we will introduce you to AI agents and the fundamental concepts behind agentic systems. We will explore frameworks used to build these systems and examine their potential applications in solving complex tasks autonomously. This module will set the stage for building more sophisticated AI-driven solutions in the following lessons.

Agentic Systems – crewAI

In this module, we will focus on the crewAI framework, where you’ll learn how to work with agents to build powerful AI systems. We’ll guide you through the process of setting up a crewAI project, defining tasks, and debugging agent workflows. Additionally, you will extend these systems by integrating custom tools and ensuring smooth execution through testing.

Agentic Systems – AG2

In this module, we will dive into AG2, a powerful framework for building conversational AI agents. You will learn to code systems with multiple agents interacting with each other and with humans. Additionally, we’ll explore how to integrate external tools to extend the functionality of your agents and create more dynamic and adaptable AI systems.

Agentic Systems – OpenAI Agents SDK

In this module, we will explore the OpenAI Agents SDK and its features for building complex AI systems. You’ll learn how to create workflows that handle agent handoffs and ensure smooth operation. The course will also cover essential techniques for applying guardrails, ensuring safe agent behavior, and using tracing for debugging and performance monitoring.

Agentic Systems – Google ADK

In this module, we will introduce the Google Agent Development Kit (ADK) and guide you through building multi-agent systems. You will learn to work with function tools to extend agent capabilities and tackle complex tasks. This will enhance your ability to design sophisticated agent-driven workflows with the ADK framework.

Agent Interactions (MCP, A2A, ACP)

In this module, we will focus on agent-to-agent communication protocols like MCP, A2A, and ACP. You will gain hands-on experience in setting up and testing MCP server-client interactions to facilitate effective communication between agents. This will equip you with the skills to build more dynamic and interconnected agent systems.

Model Finetuning

In this module, we will introduce you to model finetuning techniques, focusing on methods like LoRA. You’ll learn how to adapt pre-trained models to specific tasks and fine-tune their performance for better results. This skill will be crucial for optimizing AI models to meet the needs of different applications.