FEEDNET Artificial Intelligence Feednet Teach

Course syllabus

Generative AI with Agentic AI

Build intelligent applications. Create AI agents. Automate workflows.

Duration 4 months Format 100% online · live · interactive · flexible Fee ₹19,999 (was ₹79,988 · 75% off)

Relaunch offer — first 2 batches only

4-month live online programme. Python, large language models, RAG, LangChain, LangGraph, multi-agent systems, and a production capstone.

Who it’s for: Software, cloud, RPA, PM, data, finance/ops/HR professionals who want to build LLM agents with tools, APIs, and memory.

Pay & enroll · ₹19,999 Download brochure Enquire WhatsApp

Highlights

What this program covers

  • 4-month live programme
  • LLMs, prompt engineering, RAG, and vector search
  • LangChain, LangGraph, MCP, and guardrails
  • Capstone: enterprise agentic AI workflow platform

Opening slide

GENERATIVE AI with AGENTIC AI

Generative AI with Agentic AI: Prompt, LLM, RAG, Tools, Memory, Agent, Guardrails.
  1. PromptClear instructions — few-shot, CoT, structured output.
  2. LLMGPT, Claude, Gemini, open models — sampling and APIs.
  3. RAGChunk, embed, retrieve — answers grounded in your docs.
  4. ToolsFunction calling, APIs, files, search, MCP.
  5. MemoryWhat the agent should remember across steps.
  6. AgentPlan, act, observe — loop until the goal is done.
  7. GuardrailsHITL, injection defense, logs, cost, deploy.

Whiteboard

Diagrams we draw in week one

Keep these three pictures. Every later module (RAG, LangGraph, MCP, guardrails) is a sharper version of one of them.

01 — pick the right pattern

Chatbot vs RAG vs Agent

Three columns: a chatbot is a one-shot LLM reply; RAG retrieves documents then answers; an agent plans, calls tools, stores memory, and may wait for a human before acting.
  1. ChatbotUser → LLM → reply. One shot. No tools.
  2. RAGQuestion → vector store → LLM + chunks → grounded answer.
  3. AgentGoal → plan → tools/MCP → memory → HITL → action. Loops until done.

Most failed “agent” projects were a chatbot that should have been RAG — or RAG that needed a real loop. We start here so you do not overbuild.

02 — the inner loop

ReAct

ReAct cycle: Thought, Action or tool call, Observation, back to Thought. Human-in-the-loop when the stake is high.

Thought → tool → observation → thought. LangChain and LangGraph run a version of this loop.

03 — a crew, not a blob

Multi-agent

Orchestrator delegates to Researcher, Writer, and Reviewer agents that share memory, then a human approves before deploy.

LangGraph graphs: specialists, shared memory, then a human gate before anything ships.

  1. 01Python
  2. 02LLMs
  3. 03RAG
  4. 04Agents
  5. 05LangGraph
  6. 06MCP
  7. 07Guardrails
  8. 08Capstone

After you pay

What enrollment unlocks

Payment is processed on Feednet Solutions. Approved students get the learning home, recordings, and assessments there.

  • Access to everyday class recording sessions
  • Daily assessments + module-wise and weekly assessments
  • Saturday assignments that raise your score
  • Monthly grand test / monthly offline or online interview
  • Learning-home (LMS) dashboard after enrollment
  • Resume, portfolio & mock-interview support on the paid track

Curriculum

Modules

  1. Module 1 — Python for AI application development
    • Python programming, functions, modules, OOP and virtual environments
    • APIs, JSON, async basics, logging, exceptions and package management
  2. Module 2 — AI, machine learning and deep learning foundations
    • AI vs. ML vs. deep learning; neural networks and model training concepts
    • How foundation models and transformer-based LLMs work
  3. Module 3 — Generative AI and large language models
    • LLM capabilities, tokens, context windows and model limitations
    • Model selection, temperature, sampling and inference concepts
  4. Module 4 — Prompt engineering and structured outputs
    • Instruction design, roles, context, examples and prompt patterns
    • Few-shot prompting, structured JSON outputs and prompt evaluation
  5. Module 5 — LLM APIs and application integration
    • Integrating hosted LLM APIs into Python applications
    • Streaming, conversation history, function calling and error handling
  6. Module 6 — Embeddings and vector databases
    • Text embeddings, semantic similarity and vector search
    • Chunking strategies, metadata filtering and retrieval fundamentals
  7. Module 7 — Retrieval-augmented generation (RAG)
    • Document ingestion, parsing, chunking, indexing and retrieval
    • Grounded responses, citations, hybrid search and retrieval evaluation
  8. Module 8 — Advanced RAG and knowledge systems
    • Query rewriting, reranking, metadata filters and hybrid retrieval
    • RAG evaluation, hallucination reduction and knowledge-base updates
  9. Module 9 — AI agents and tool use
    • Agent architecture, reasoning loops, planning and task execution
    • Tools, function calling, APIs, memory and human approval
  10. Module 10 — LangChain and LangGraph
    • Chains, prompts, retrievers, tools and application components
    • Stateful graphs, nodes, edges, checkpoints and controlled workflows
  11. Module 11 — Multi-agent systems and collaboration
    • Agent roles, delegation, handoffs and shared task context
    • Multi-agent orchestration patterns and workflow coordination
  12. Module 12 — Model Context Protocol (MCP) and integrations
    • MCP concepts, tools, resources and client-server architecture
    • Connecting AI applications with external systems and services
  13. Module 13 — AI evaluation, safety and guardrails
    • Testing prompts, RAG pipelines, agents and tool execution
    • Groundedness, quality checks, guardrails, privacy and access control
  14. Module 14 — Deployment, monitoring and optimization
    • FastAPI deployment, Docker and environment configuration
    • Tracing, logging, latency, token cost and production monitoring
  15. Module 15 — Advanced agentic AI design and interview preparation
    • Agent reliability, failure recovery, observability and design trade-offs
    • Architecture discussions, project walkthroughs and scenario-based interviews
  16. Projects and final capstone
    • AI document assistant (RAG chatbot)
    • Customer support AI agent
    • Multi-agent research and report automation
    • Autonomous AI coding and debugging agent
    • Capstone: enterprise agentic AI workflow platform — RAG, tools, agents, approvals and evaluation

On completion

Certificate of completion

After you finish Generative AI with Agentic AI, Feednet Teach issues a certificate like this in your name. The learner name is hidden on this sample.

Sample
Feednet Solutions Pvt Ltd Hyderabad, Telangana, India CIN: U72900TG2019PTC134896

Certificate of Completion

Generative AI with Agentic AI

This certificate is proudly presented to

for successfully completing the Generative AI with Agentic AI training program, demonstrating dedication and proficiency in large language models, retrieval-augmented generation, LangChain, LangGraph, and agentic AI workflows.

Duration: 4 Months

Program Director PROGRAM DIRECTOR
Authorized Signatory AUTHORIZED SIGNATORY

Certificate ID: FSPL-GAI-SAMPLE · Issued on completion

Sample only. Your name, dates, and certificate ID appear after you complete Generative AI with Agentic AI.
Pay Enquire WhatsApp
WhatsApp