🤖

Gen AI & Agentic AI

Generative AI, LLMs, RAG & Autonomous Agent Systems

Free Python + Linux for Multi-Cloud DevOps Students
New Batch Starting Soon!
Free Python + Linux for Multi-Cloud DevOps Students
New Batch Starting Soon!
🤖

Gen AI & Agentic AI Syllabus

From LLM fundamentals to production multi-agent systems

1
AI Foundations

Foundations of AI, ML & Generative AI

  • AI vs Machine Learning vs Deep Learning
  • Supervised, Unsupervised & Reinforcement Learning basics
  • Neural Networks and how they learn
  • What makes Generative AI different
  • Discriminative vs Generative models
  • Real-world GenAI applications across industries
2
Python for AI

Python & Data Foundations for GenAI

  • Python essentials for AI engineering
  • NumPy, Pandas for data handling
  • Working with APIs and JSON
  • Virtual environments and dependency management
  • Jupyter notebooks & Google Colab
  • Intro to PyTorch tensors
3
Transformer Architecture

Understanding Transformers & LLMs

  • Attention is All You Need — self-attention mechanism
  • Transformer encoder-decoder architecture
  • Tokenization strategies (BPE, WordPiece, SentencePiece)
  • Embeddings and positional encoding
  • Transformer decoders & MoE (Mixture of Experts) routing
  • How GPT, Llama, Claude, Gemini architectures differ
4
Large Language Models

Working with LLMs & Model APIs

  • OpenAI API, Anthropic Claude API, Google Gemini API
  • Model selection: context window, latency, cost trade-offs
  • Open-source LLMs via Hugging Face & Ollama
  • Streaming responses and function/tool calling
  • Structured outputs & JSON mode
  • Rate limits, retries and cost optimization
5
Prompt & Context Engineering

Prompt Engineering & Context Engineering (2026)

  • Zero-shot, few-shot and chain-of-thought prompting
  • Role-based and system prompting
  • Context engineering: curating everything the model sees, not just what you say
  • Context windows, token budgets and context compression
  • Memory, retrieved knowledge, tool results & state as context
  • Prompt injection risks, guardrails and evaluation
6
Embeddings & Vector Search

Embeddings, Vector Databases & Search

  • Text embeddings and semantic similarity
  • Vector databases: Pinecone, Chroma, Weaviate, FAISS, Milvus
  • Indexing strategies and ANN search (HNSW, IVF)
  • Chunking strategies for documents
  • Hybrid search (keyword + vector)
  • Vector search optimization and evaluation
7
RAG Systems

Retrieval-Augmented Generation (RAG)

  • RAG architecture: retrieve, augment, generate
  • Building a document Q&A pipeline end-to-end
  • Advanced RAG: re-ranking, query rewriting, HyDE
  • Multi-source and multi-modal RAG
  • Evaluation metrics: RAGAS, BERTScore, faithfulness
  • Production RAG pitfalls and best practices
8
Fine-Tuning & Customization

Fine-Tuning & Model Customization

  • When to fine-tune vs prompt vs RAG
  • Parameter-efficient fine-tuning: LoRA and QLoRA
  • Instruction tuning and RLHF basics
  • Dataset preparation for fine-tuning
  • Fine-tuning on Hugging Face & cloud platforms
  • Evaluating fine-tuned models
9
Orchestration Frameworks

LangChain, LlamaIndex & Orchestration

  • LangChain chains, tools and memory
  • LlamaIndex for data-centric LLM apps
  • Building conversational memory (buffer, summary, vector)
  • Output parsers and structured generation
  • Connecting LLMs to external tools and APIs
  • Debugging and tracing with LangSmith
10
Model Context Protocol (MCP)

MCP: The Standard for AI Tool & Agent Connectivity

  • What is MCP and why it became the 2026 industry standard
  • MCP architecture: hosts, clients and servers
  • Core primitives: tools, resources and prompts
  • Transport layers: stdio, SSE and streamable HTTP
  • Building an MCP server with the Python/TypeScript SDK
  • Connecting Claude, ChatGPT & IDE agents to MCP servers (GitHub, DBs, APIs)
11
Agentic AI Design

Agentic AI: Design Patterns & Reasoning

  • What is Agentic AI — beyond simple Q&A
  • ReAct pattern: reasoning + acting
  • Plan-and-Execute agent architecture
  • Reflection and self-critique loops
  • Tool use: code execution, web search, APIs, MCP tools
  • Agent memory: short-term vs long-term
12
Multi-Agent Orchestration

Multi-Agent Systems & Orchestration

  • LangGraph for stateful agent graphs
  • CrewAI for role-based multi-agent teams
  • AutoGen for conversational multi-agent workflows
  • Agent-to-agent (A2A) communication protocols
  • Supervisor/orchestrator agent patterns
  • Designing agents that plan, execute, reflect and iterate
13
AI Coding Assistants

AI-Powered Development & Copilots

  • GitHub Copilot for code generation
  • Claude Code / Cursor / AI IDE agent workflows
  • AI-assisted code review and refactoring
  • Generating tests and documentation with AI
  • Prompting for infrastructure-as-code (Terraform/YAML)
  • AI pair-programming best practices
14
AIOps for DevOps

AI for DevOps, Monitoring & Incident Response

  • Log summarization and anomaly detection with LLMs
  • AI-powered alert noise reduction
  • Incident timeline generation and root-cause analysis
  • KubeGPT-style Kubernetes troubleshooting agents
  • LLM-powered CI/CD pipeline failure explanation
  • AI-assisted rollback and release-note generation
15
Multimodal & Diffusion Models

Multimodal AI & Diffusion Models

  • Text-to-image generation (diffusion model basics)
  • Vision-language models (GPT Vision, Gemini Vision, Claude Vision)
  • Text-to-speech and speech-to-text pipelines
  • Multimodal RAG (text + images + tables)
  • Video and audio generation overview
  • Practical use cases across content and support workflows
16
Evaluation, Safety & Governance

Evaluation, Safety & Responsible AI

  • LLM & agent evaluation metrics and benchmarks
  • Hallucination detection and mitigation
  • Bias, fairness and content-safety guardrails
  • Prompt injection, jailbreak and MCP security defenses
  • Human-in-the-loop evaluation workflows
  • Responsible AI & emerging AI-agent governance standards
17
Deployment & LLMOps

Deployment, Scaling & LLMOps

  • Deploying LLM/agent apps with FastAPI/Flask
  • Containerizing GenAI apps with Docker
  • Serving open-source models (vLLM, TGI, Ollama)
  • Caching, batching and cost control at scale
  • Observability for LLM apps: tracing, logging, cost dashboards
  • CI/CD for AI applications on AWS/Azure/GCP
18
Capstone Projects

Capstone: Real-World GenAI & Agentic Projects

  • Enterprise document Q&A chatbot with RAG
  • Autonomous multi-agent research assistant (LangGraph/CrewAI)
  • Custom MCP server + agent integration project
  • AI DevOps assistant: log analyzer + K8s troubleshooting agent
  • Customer-support agent with tool use and memory
  • End-to-end LLMOps pipeline: build, evaluate, deploy, monitor