Book Free Demo
  1. Home
  2. Courses
  3. Agentic AI Syllabus
🔥 2025 Most In-Demand

Agentic AI —Build Autonomous AI Systems

Learn to design and build AI agents that can plan, use tools, collaborate, and autonomously complete complex multi-step tasks —the next frontier of AI development.

45Hours
7Modules
6+Agents Built
✓Certificate

What Makes an AI Agent?

🧠 Reasoning

Plans multi-step tasks and decides what to do next

🔧 Tool Use

Calls APIs, searches the web, runs code, reads files

💾 Memory

Remembers context — short-term, long-term, episodic

🤝 Collaboration

Works with other agents in multi-agent systems

Frameworks & Tools

🤖 AutoGen (Microsoft) ⚓ CrewAI 🗺 LangGraph 🔷 Semantic Kernel 🐍 Python ⚡ OpenAI Assistants API 🔌 MCP (Model Context Protocol) 🛠 Function Calling ☁ Azure AI Foundry 🏠 Ollama 🧰 Tavily Search API 📊 E2B Code Interpreter

📋 Detailed Course Syllabus

Module 1: Introduction to AI Agents & Agentic Systems 5 Hours
  • What are AI Agents? Difference from chatbots and standard LLM calls
  • The agent loop: Perceive →Reason →Act →Observe →Repeat
  • ReAct (Reasoning + Acting) pattern —the foundation of agents
  • LLM as the "brain" —planning, reflection, and self-correction
  • Types of agents: Reflex, Goal-based, Utility-based, Learning agents
  • Overview of leading agentic frameworks: AutoGen, CrewAI, LangGraph, Swarm
Module 2: Tool Use & Function Calling 6 Hours
  • OpenAI Function Calling API —defining tools as JSON schemas
  • Building custom tools: web search, calculator, database query, file read/write
  • Integrating external APIs as agent tools (weather, news, e-commerce)
  • Tool selection strategy —how the agent decides which tool to call
  • Parallel tool calls —calling multiple tools simultaneously
  • Tool error handling and retry logic within the agent loop
  • Model Context Protocol (MCP) —the emerging standard for tool integration
Module 3: Agent Memory Systems 5 Hours
  • Types of agent memory: Working (short-term), Episodic, Semantic, Procedural
  • Conversation buffer and window memory implementation
  • Summary memory —compressing long conversations for context efficiency
  • Long-term memory with vector stores: storing and retrieving past interactions
  • Entity memory —tracking people, places, and objects across sessions
  • Memory in CrewAI: short-term, long-term, entity, and contextual memory
  • Practical: Building an agent that remembers user preferences across sessions
Module 4: Multi-Agent Systems with AutoGen & CrewAI 8 Hours
  • Why multi-agent systems? —Specialization, parallelism, peer review
  • AutoGen deep dive: AssistantAgent, UserProxyAgent, GroupChat patterns
  • AutoGen: Code execution, human-in-the-loop, agent conversation flows
  • CrewAI framework: Crew, Agents, Tasks, Tools, Process types
  • CrewAI: Sequential vs. Hierarchical process, manager agent patterns
  • Agent roles and persona design: Researcher, Analyst, Writer, Critic agents
  • Inter-agent communication patterns: message passing, shared state
  • Project: Build a research + report writing multi-agent system
Module 5: LangGraph —Stateful Agentic Workflows 8 Hours
  • Why LangGraph? Overcoming linear chain limitations with cyclic graphs
  • Core concepts: Nodes, Edges, State, Conditional edges, Cycles
  • State management: TypedDict state schema, reducers, annotations
  • Human-in-the-loop workflows: interrupts, checkpoints, approval steps
  • Persistence with LangGraph checkpointing (SQLite, PostgreSQL)
  • Subgraphs and nested agents for complex hierarchical workflows
  • LangGraph Studio —visual debugging and workflow monitoring
  • Project: Customer support agent with escalation and memory using LangGraph
Module 6: OpenAI Assistants API & Semantic Kernel Agents 7 Hours
  • OpenAI Assistants API: Threads, Messages, Runs, and Tools
  • File Search (built-in RAG) and Code Interpreter tool in Assistants
  • Streaming Runs for real-time agent responses
  • Microsoft Semantic Kernel —agents for .NET and Python developers
  • Semantic Kernel: Plugins, Planners, Auto-function invocation
  • Kernel Function calling —declarative AI orchestration
  • Building .NET-based AI agents with Semantic Kernel Process Framework
Module 7: Production Agents —Evaluation, Safety & Deployment 6 Hours
  • Agent evaluation: measuring task success, trajectory correctness, tool use
  • Guardrails for agents: preventing harmful actions, scope limiting
  • Prompt injection defense in agentic systems
  • Reliability patterns: retry, fallback, timeout, circuit breaker
  • Cost management: monitoring token usage in multi-step agent runs
  • Deploying agents as APIs on Azure Container Apps
  • Capstone: Full autonomous AI research agent —given a topic, produces a comprehensive report, searches the web, writes code, and sends an email summary

🤖 Be Among the First Agentic AI Developers in India

Agentic AI is the fastest-growing skill in 2025. Get ahead of the curve —book your free demo today.

Book FREE Demo Class WhatsApp Us

Start Your IT Career Today - Book a FREE Demo Class!

Azure, .NET, Java, Python, SQL Server - Online & Classroom Training with Placement Support

Enquire Now WhatsApp Us
;