Capability Development / AI
Generative AI & AI Agents (Multi Domain)
This program is a comprehensive, hands-on learning journey into Generative AI and AI Agents, covering foundational concepts, prompt engineering, RAG, cloud-based AI services, multimodal models, fine-tuning, agent development, monitoring and responsible AI. Learners explore Amazon Bedrock, Azure AI Foundry, Google AI Studio and Vertex AI while building real solutions across text generation, image generation, retrieval, automation, evaluation and enterprise deployment.
Program Overview
This program is a comprehensive, hands-on learning journey into Generative AI and AI Agents, covering foundational concepts, prompt engineering, RAG, cloud-based AI services, multimodal models, fine-tuning, agent development, monitoring and responsible AI. Learners explore Amazon Bedrock, Azure AI Foundry, Google AI Studio and Vertex AI while building real solutions across text generation, image generation, retrieval, automation, evaluation and enterprise deployment.
Audience Profile
For developers with proficiency in Python, familiarity with Jupyter and VS Code, a basic understanding of machine learning and NLP, and working knowledge of REST APIs and Git, who want to build and deploy generative AI and multi-agent solutions across AWS, Azure and Google Cloud.
Course Syllabus
Module 1: Foundations of Generative AI
Topics
- Applications, evolution and architecture of Generative AI
- Foundation models and the transformer architecture
- Text, image, audio and multimodal generation
- Open-source vs. closed-source LLMs (LLaMA, Mistral, Claude, GPT)
Labs / Practice
- Load and infer with Hugging Face models
Module 2: Prompt Engineering, RAG and LangChain
Topics
- Zero-shot, few-shot and chain-of-thought prompting
- Role prompting, formatting and avoiding hallucinations
- RAG architecture, vector databases (FAISS, Chroma, Weaviate)
- LangChain and LlamaIndex overview
Labs / Practice
- Build complex prompts for industry use cases
- Build a RAG chatbot with OpenAI, FAISS and LangChain
Module 3: Generative AI on AWS, Azure and Google Cloud
Topics
- Leading foundation models: Titan, Claude, Command, Llama, Gemini
- Amazon Bedrock, Amazon Q and SageMaker for Generative AI
- Google AI Studio, Gemini API and Vertex AI Studio
- Azure AI Foundry projects, hubs, services and model catalog
Labs / Practice
- Experiment across Bedrock, Vertex AI and Azure AI Foundry playgrounds
Module 4: Text Generation and RAG Across Clouds
Topics
- Text generation, summarization, QnA and entity extraction on Bedrock
- Azure AI Foundry chat completions and text embeddings
- Managed RAG, Azure AI Search and Vertex AI RAG Engine
Labs / Practice
- Build a RAG-powered application on each cloud
Module 5: Image, Video and Multimodal Generation
Topics
- Bedrock Titan Image Generator and Amazon Nova
- Imagen 3 and Veo 2/3 in Google AI Studio
- Azure AI Foundry image generation and DALL·E 3
Labs / Practice
- Generate images and video across cloud platforms
Module 6: Fine-Tuning and Model Customization
Topics
- Fine-tuning options in Amazon Bedrock and Google AI Studio
- Fine-tuning with Azure AI Foundry
- LoRA, QLoRA and parameter-efficient fine-tuning (PEFT)
Labs / Practice
- Fine-tune and invoke a custom model
Module 7: AI Agents Across Platforms
Topics
- Agents for Amazon Bedrock, Gemini API and Azure AI
- LangChain and LangGraph agent workflows
- Google Agent Development Kit, A2A and Model Context Protocol
- Multi-agent systems: LangGraph, CrewAI, OpenAI Agents SDK, Semantic Kernel
Labs / Practice
- Build and deploy AI agents across AWS, Azure and Google Cloud
Module 8: Monitoring, Ethics and Developer Tools
Topics
- Observability, tracing and LLM evaluation (MLflow, Evidently AI)
- Responsible AI, guardrails and LLMOps
- Developer AI tools: GitHub Copilot, Amazon Q Developer, Cursor, Claude Code
Labs / Practice
- Evaluate LLM outputs and set up an observability dashboard