Home Courses Agentic AI
Data Science & AI LangGraph · CrewAI · AutoGen · MCP ✔ 3 Levels Available

Agentic AI
Full Programme

A comprehensive hands-on curriculum engineered for AI engineers and enterprise architects. Master single-agent cognitive patterns (ReAct, Plan-and-Solve), stateful graph orchestration (LangGraph), multi-agent swarms (CrewAI, AutoGen), Model Context Protocol (MCP), and production LLMOps with governance and Human-in-the-Loop workflows.

🎯
Format
Theory + Lab/Assignments
🏆
Certificate
Arich Certified
Enrol Now →
Course Code
AGENT-AI
Click a Level to View its Syllabus
L1
AGENT-L1 · Associate
6 Modules · Foundations & Architecture
›
L2
AGENT-L2 · Professional
4 Modules · Frameworks & Swarms
›
L3
AGENT-L3 · Architect
4 Modules · MCP & Enterprise Ops
›
📄 Agentic AI — Associate Level (AGENT-L1)
L1
AGENT-L1
Agentic Foundations & Core Paradigms
01
Agent Concepts & Single-Agent Architectures
▼
1.1 Autonomous Agent Architecture & Cognitive Reasoning
ReAct Pattern (Reasoning + Acting) & Execution Cycles
Plan-and-Solve Strategies & Self-Reflection Mechanisms
Memory Systems: Short-term Working Memory & Long-term Persistent Memory
Tool Binding, Structured Outputs & Native Function Calling
Hands-on Lab: Build a ReAct Agent from scratch in Python with Memory & Web Search Tool Execution
02
Agentic Frameworks (LangGraph, AutoGen, CrewAI)
▼
2.1 Multi-Agent Frameworks & Stateful Graph Foundations
LangGraph State Graphs, Nodes, Edges & Cyclical Execution
AutoGen Conversational & Group Chat Multi-Agent Architectures
CrewAI Hierarchical Tasks, Role-Playing Agents & Sequential Processes
LlamaIndex Agentic Workflows & Query Routing Engines
Hands-on Lab: Develop Stateful Multi-Agent Research Assistant using LangGraph with Conditional Branching
03
Multi-Agent Swarms & Communication Protocols
▼
3.1 Hierarchical & Peer-to-Peer Swarms Architecture
Hierarchical Agent Teams & Supervisor-Worker Patterns
Leader-Worker & Blackboard Collaborative Architectures
Agent Communication Protocols: JSON RPC & Agent-to-Agent (A2A) Messaging
Consensus Mechanisms, Conflict Resolution & Collaborative Planning
Hands-on Lab: Construct Multi-Agent Software Engineering Team (Product Manager, Coder, Reviewer) using CrewAI
04
Enterprise Tool & API Orchestration
▼
4.1 Enterprise Tool Integration & Dynamic OpenAPI Tool Ingestion
Model Context Protocol (MCP) Architecture, Clients & Servers
Database Query & Safe Mutation Execution with Guardrails
Safe Code Interpreters & Sandboxed Code Execution (E2B / Docker)
Human-in-the-Loop (HITL) Interactive Approval Workflows
Hands-on Lab: Implement Enterprise Agent with Model Context Protocol (MCP) and HITL Approval Workflows for Database Mutations
05
Agentic Memory, RAG Integration & Knowledge Grounding
▼
Vector Databases (Chroma, Pinecone, Qdrant) for Persistent Agent Memory
Episodic, Semantic, and Procedural Memory Retrieval Pipelines
Hybrid Search (BM25 + Dense Vectors) for Domain Knowledge Grounding
Context Window Compaction, Summarization & Token Budgeting
Hallucination Mitigation via Grounded Reflection & Source Attribution
Hands-on Lab: Build an Agent with Long-Term Episodic Memory & Dynamic Knowledge Retrieval
06
Production Agentic Operations, Observability & Evaluation
▼
Agent Telemetry, Tracing & Observability with LangSmith & Phoenix
Evaluation Metrics & Automated Benchmark Testing (Ragas, TruLens, DeepEval)
Guardrails, Prompt Injection Defense & Safety Filters (NeMo, Llama Guard)
Packaging Agents as Asynchronous Microservices with FastAPI & WebSockets
Cross-Cloud Deployment Architecture (AWS SAP / Azure AI / GCP PCA standards)
Capstone Project: End-to-End Enterprise Agent Deployment with Tracing, Automated Evals & Production Guardrails
📄 Agentic AI — Professional Level (AGENT-L2)
L2
AGENT-L2
Multi-Agent Systems & Framework Specialist
01
Stateful Graph Orchestration with LangGraph
▼
State Schemas, Reducers & Channel-Based State Passing in LangGraph
Conditional Routing, Cyclical Execution Loops & Breakpoints
Persistence Layers: Checkpointers (SQLite, Postgres) for Resumable Agent Sessions
Time-Travel Debugging & Manual State Mutation
Hands-on Lab: Build an Autonomous Bug-Fixing Agent Graph with Checkpointing & Verification Nodes
02
Multi-Agent Collaboration with CrewAI & AutoGen
▼
CrewAI Custom Agents, Tools, Tasks & Hierarchical Manager Workflows
AutoGen ConversableAgent, UserProxyAgent & GroupChatManager Patterns
Dynamic Speaker Selection & Round-Robin Conversation Orchestration
Integrating Custom Python Tools & Web Scrapers into Agent Workflows
Hands-on Lab: Develop a Competitive Intelligence Multi-Agent Team with Real-Time Web Scraping & Synthesis
03
Distributed Swarms & Agent Communication Protocols
▼
Decentralized Multi-Agent Swarms & Peer-to-Peer Task Delegation
Agent Communication Protocols: JSON-RPC, WebSocket Streaming & Message Queues (RabbitMQ/Kafka)
Consensus Algorithms: Voting, Delphi Methods & Debate-Based Decision Making
Deadlock Prevention & Loop Detection in Autonomous Multi-Agent Networks
Hands-on Lab: Implement a Financial Market Simulation with Autonomous Buyer, Seller & Auditor Agents
04
Autonomous Multi-Agent SDLC Project
▼
Architecture: Product Manager Agent -> Lead Architect -> Senior Coder -> QA Engineer -> Security Auditor
Automated Code Generation, Test Execution, Code Review & Git PR Automation
Dynamic Error Rectification via Feedback Loops between QA and Coder Agents
Hands-on Lab: Build an Autonomous Feature-Development Swarm that Takes JIRA Issues to Merged PRs
📄 Agentic AI — Architect Level (AGENT-L3)
L3
AGENT-L3
Enterprise Agent Architecture, MCP & Governance
01
Model Context Protocol (MCP) & Enterprise Tool Ingestion
▼
Anthropic Model Context Protocol (MCP) Specification, Architecture & Client-Server Lifecycle
Building Custom MCP Servers for Internal Databases, Enterprise APIs & File Systems
Dynamic Tool Discovery, Schema Validation & Capability Negotiation
Cross-Cloud MCP Interoperability (AWS Bedrock, Azure OpenAI, GCP Vertex AI)
Hands-on Lab: Construct an Enterprise MCP Server Connecting SQL Warehouses, CRM APIs & Cloud Storage
02
Human-in-the-Loop (HITL) Governance & Guardrails
▼
Asynchronous Human Approval Queues for High-Stakes Tool Executions
Enterprise Guardrails: NeMo Guardrails, Llama Guard & Custom Semantic Filters
Prompt Injection, Jailbreak Defense & Tool Hijacking Prevention Strategies
Audit Logging, Compliance Tracking & Cryptographic Tool Verification
Hands-on Lab: Implement a Financial Transaction Agent with Dual-Control HITL Approval & Guardrails
03
Scalable Cognitive Memory & Graph RAG
▼
Hierarchical Episodic, Semantic, and Procedural Memory Systems
Graph-RAG Integration: Combining Knowledge Graphs (Neo4j) with Vector Search
Dynamic Context Compaction, Reflection Trees & Long-Horizon Planning
Sub-Goal Decomposition & Tree-of-Thought (ToT) Agentic Reasoning
Hands-on Lab: Build an Enterprise Knowledge Graph Agent with Self-Correcting Graph-RAG Pipelines
04
Production AgentOps, Tracing & Automated Evaluation
▼
Full-Stack Observability: OpenTelemetry, LangSmith, Phoenix & Arize Telemetry
Automated Agentic Evaluation Frameworks: Ragas, TruLens, DeepEval Benchmarking
Latency, Cost & Token Optimization: Model Routing & Prompt Caching
Enterprise Deployment: Docker, Kubernetes, Helm Charts & Production Microservices
Capstone Project: Production Deployment of a Multi-Agent Enterprise Customer Operations Platform