Agentic AI: Build AI Agents with MCP, FastMCP & Pydantic AI
Published 7/2026
Created by Nandakishor M
MP4 | Video: h264, 1280×720 | Audio: AAC, 44.1 KHz, 2 Ch
Level: All Levels | Genre: eLearning | Language: English | Duration: 10 Lectures ( 1h 23m ) | Size: 787.7 MB
Learn JSON-RPC and the Model Context Protocol, build multi-agent systems, run local LLMs, and ship a JWT-secured agent
What you’ll learn
⚡ Build AI agents from first principles using the Model Context Protocol (MCP), understanding the JSON-RPC loop behind every agent.
⚡ Turn any Python function into an agent tool in minutes with FastMCP, and orchestrate multi-agent, multi-tool systems using Pydantic AI.
⚡ Run large language models locally with Ollama and GGUF quantization, so your agents work offline with no API bills.
⚡ Deploy a working AI agent as a production REST API secured with JWT authentication using FastAPI.
⚡ Understand LLM internals that matter for agents: tokens, embeddings, attention, FP16 and quantization.
Requirements
❗ Comfortable writing basic Python (functions, classes, pip). No AI or machine learning background needed, everything about agents and LLMs is taught from scratch.
❗ A computer that can run Python 3. A GPU is helpful for local models but not required.
Description
Stop chatting with AI. Start building agents that do things.
Anyone can talk to a language model. Very few people can build an agent that reads files, calls tools, coordinates other agents, and runs as a real service you can deploy. This course takes you all the way there, from first principles to a secured, production-ready AI agent, with code you write and run yourself. No hand-waving, and no copy-paste magic you do not understand.
You will learn
✅ The tool-calling loop that turns a plain language model into an agent that takes action
✅ The Model Context Protocol (MCP) explained properly: it is really just JSON-RPC 2.0 and three method names, and you will read and write it by hand
✅ FastMCP, so a single Python function becomes a fully described agent tool in about ten lines
✅ Your first real agent with Pydantic AI that writes a file to disk on your own machine
✅ Multi-agent, multi-tool systems where a manager agent delegates to specialist agents, with safety checks on risky actions
✅ What is really inside an LLM: tokens, embeddings, attention, FP16, and GGUF quantization, taught visually and from scratch
✅ How to run models locally with Ollama, fully offline, with no API keys and no cloud bills
✅ A capstone where you deploy your agent as a FastAPI REST API secured with JWT authentication, and trace the full request flow
Most courses stop at “here is a prompt.” This one ends with a deployed, authenticated agent API, the part almost no one teaches. Every concept is grounded in code you write and run, so you finish with working projects, not just notes.
This course is taught by an AI researcher and founder who builds these systems for a living, with published work in applied AI. You should be comfortable writing basic Python. You do not need any AI or machine learning background, because everything about agents, MCP, and LLMs is taught from the ground up.
Enroll now and build your first production AI agent.
Who this course is for
⭐ Python developers who want to build real AI agents, not just chat with them, and ship them to production.
⭐ Engineers and students curious about MCP, FastMCP and agentic AI who prefer building from first principles over copy-paste tutorials.
⭐ Anyone who wants to run LLMs locally and deploy a secured agent API without relying on paid cloud APIs.
AEGNATRFDEBELDHERWITAFEMCPERFATSRF

you must be registered member to see linkes Register Now