Keerti Purswani - Hands-On Generative AI

File Name:	Keerti Purswani - Hands-On Generative AI
Content Source:	https://educosys.com/course/hands-on-generative-ai
Genre / Category:	Courses for Developers
Language:	ENGLISH
Original Price:	₹29,999

Description:

Learn generative artificial intelligence from scratch to an advanced level. Master creating, training, fine-tuning, and deploying modern AI models, developing intelligent agents, and creating comprehensive applications based on large language models (LLM).

The course combines the theoretical foundations of machine learning with practical development. You will study neural networks, generative models, transformer architecture, RAG technologies, vector databases, and modern AI development tools, including TensorFlow, Hugging Face, LangChain, LangGraph, and Ollama.

Practical assignments are given special attention. You will independently implement a transformer from scratch, train generative models, master fine-tuning LLMs, and develop your own AI applications, including an intelligent chatbot with memory, internet search, and the ability to work with external data sources.

What You Will Learn

• Understand the mathematical foundations of artificial intelligence and the principles of neural networks.

• Develop and train generative models GAN and VAE using TensorFlow.

• Understand the architecture of transformers, attention mechanisms, and the structure of large language models GPT and BERT.

• Independently implement transformers and apply pre-trained language models.

• Fine-tune LLM with LoRA, QLoRA, Hugging Face, and Unsloth.

• Create AI agents and automated workflows with LangChain and LangGraph.

• Work with vector databases and develop RAG systems.

• Create comprehensive AI applications with an interface on Streamlit.

• Apply modern technologies MCP, Ollama, Mixture of Experts, and multimodal models.

• Develop and deploy your own projects based on generative artificial intelligence.

Course Program

1. Fundamentals of Generative Artificial Intelligence

Start with the fundamental principles of artificial intelligence, the mathematical foundations of machine learning, and the operation of neural networks.

Module Topics:

• Introduction to artificial intelligence.

• Mathematical foundations of AI.

• Probability theory, statistics, and linear algebra.

• Basic principles of neural network operation.

• Gradient descent and optimization methods.

• Neural network architectures: Feedforward, RNN, and CNN.

Practical Projects:

• Create a simple neural network using TensorFlow.

• Train an autoencoder on the MNIST dataset.

2. Deep Generative Models

Explore the main architectures of generative artificial intelligence and learn to create models capable of generating new data.

Module Topics:

• Differences between discriminative and generative models.

• Generative adversarial networks (GAN).

• Variational autoencoders (VAE).

• Probabilistic data generation using VAE.

• Develop generative models in TensorFlow.

• Visualize metrics and training results with TensorBoard.

Practical Projects:

• Four mini-projects using TensorFlow.

• Implement GANs for generating handwritten digits.

• Train VAE for generating face images based on the CelebA dataset.

3. Transformers and Large Language Models (LLM)

Understand the architecture underlying modern language models. Study attention mechanisms and independently implement a transformer.

Module Topics:

• Recurrent neural networks RNN and LSTM.

• Transformer architecture.

• Attention mechanisms: Self-Attention and Positional Encoding.

• Encoder-Decoder architecture.

• Pre-training tasks: Masked Language Modeling (MLM) and Causal Language Modeling (CLM).

• Architectures and features of GPT and BERT models.

Practical Projects:

• Major project: implement a transformer from scratch.

• Text sentiment analysis using BERT.

4. Fine-tuning LLM, LangChain, and LangGraph

Learn to adapt pre-trained language models for specific tasks and create intelligent systems using modern frameworks.

Module Topics:

• Pre-training and fine-tuning language models.

• Efficient fine-tuning methods LoRA and QLoRA.

• Working with Hugging Face libraries and models.

• Adapting language models for text summarization and answering questions.

• LangChain installation and configuration.

• Main components of LangChain: Prompts, Memory, Chains, and Agents.

• LangGraph: nodes, states, StateGraph, and workflows.

• Develop AI agents.

Practical Project:

Creating a question-answering application using LangChain.

5. Vector Databases and RAG

Master Retrieval-Augmented Generation (RAG) — a technology that allows language models to use information from external sources to form responses.

Learn to work with vector databases, create RAG pipelines, and develop comprehensive AI applications with a user interface.

Module Topics:

• Principles of vector databases.

• Using ChromaDB.

• Architecture and practical application of RAG.

• Building RAG pipelines with LangChain.

• Developing user interfaces with Streamlit.

• Integrating LLM with external tools and information sources.

Practical Projects:

Project 1. Creating a comprehensive AI chatbot similar to ChatGPT.

Develop an intelligent chatbot using Streamlit, LangGraph, ChromaDB, and large language models. Implement dialogue memory, web search tool integration, and the ability to use external data.

Project 2. Multimodal AI application.

Create an application on Streamlit combining several capabilities of generative artificial intelligence:

• Image generation.

• Automatic image description creation.

• Generating textual descriptions of video materials.

6. Modern Technologies and Architectures of Generative AI

Get acquainted with current tools, protocols, and architectural approaches used in the development of modern AI systems.

Module Topics:

• Model Context Protocol (MCP): connecting AI models to external tools and data sources.

• Ollama: local deployment and use of language models.

• Fine-tuning models using Unsloth.

• Mixture of Experts (MoE) architecture.

• Chain of Thought: methods for organizing multi-step reasoning in language models.

• DeepSeek architecture.

Practical Projects:

• Fine-tuning language models using Unsloth.

• Practical application of studied tools and architectures.

7. Advanced Methods and Multimodal Artificial Intelligence

The final module is dedicated to modern approaches to the development and optimization of generative models, as well as technologies for processing various types of data.

Module Topics:

• Model distillation (Model Distillation).

• Diffusion models (Diffusion Models).

• Visual transformers (Vision Transformers).

• Multimodal artificial intelligence models.

• CLIP architecture.

• Prompt engineering.

• Practical projects using the studied technologies.

Learning Outcome

By the end of the course, you will understand not only the principles of modern generative models but the entire process of AI application development: from the mathematical foundations and training neural networks to creating intelligent agents and deploying comprehensive solutions based on LLM.

Practical projects will help consolidate the acquired knowledge and develop the skills necessary for independent development of generative models, RAG systems, AI chatbots, and multimodal applications.

File Information

Submitter S A N

Submitted 10/10/2026

Category Paid Coding Courses

Sale page https://educosys.com/course/hands-on-generative-ai

View File

is this complete?