Theme3
🚀 100 Days of Building Autonomous AI Systems
Days 1–10: Python + Environment Setup
Day 1: Install Miniconda, Python 3.11
Day 2: Create Conda environments, install basic libraries (pandas, requests)
Day 3: Setup Jupyter Notebook + VSCode
Day 4: Python Basics: Variables, Loops, Functions
Day 5: Git and GitHub Setup: First Commit
Day 6: Markdown Basics + GitHub Readmes
Day 7: Install Docker and Docker Compose
Day 8: Create your first simple Dockerfile (Python Hello World)
Day 9: Basic REST APIs using FastAPI
Day 10: LinkedIn Post: "Setting up for 100 Days of Autonomous AI 🚀"
Days 11–20: Core Machine Learning and LLM APIs
Day 11: Train/Test Split + Linear Regression model (scikit-learn)
Day 12: Intro to Transformers (Huggingface basics)
Day 13: Use OpenAI API for Text Completion
Day 14: Understand Embeddings: Generate Sentence Embeddings
Day 15: Build a basic FAISS index and search over it
Day 16: Document chunking with LangChain TextSplitter
Day 17: Create RAG (Retrieval-Augmented Generation) system
Day 18: Build a FastAPI endpoint serving your RAG
Day 19: Dockerize your FastAPI+RAG app
Day 20: Post your first GenAI API on LinkedIn
Days 21–30: LangChain and Advanced RAG
Day 21: LangChain Prompt Templates
Day 22: Chains: SequentialChain and LLMChain
Day 23: Build a Memory Chatbot (ConversationBufferMemory)
Day 24: Tool Creation in LangChain
Day 25: Building a LangChain Agent using Tools
Day 26: LangChain + FAISS end-to-end Pipeline
Day 27: Streamlit UI for your chatbot
Day 28: Vector Store Alternatives: Chroma, Qdrant
Day 29: Deploy LangChain RAG on Render or Railway
Day 30: LinkedIn Post: "Building Smarter GenAI Apps with LangChain"
Days 31–40: Prefect and Workflow Automation
Day 31: Install Prefect, create your first Flow
Day 32: Build a Scheduled Flow (daily document updater)
Day 33: Prefect Work Pools and Deployments Basics
Day 34: Handling Failures and Retries in Prefect
Day 35: Create a GenAI Content Pipeline (Extract -> Summarize -> Post)
Day 36: Prefect Deployment with Docker
Day 37: Environment Variables and Secrets in Prefect
Day 38: Prefect Orchestration of GenAI Chains
Day 39: Prefect Monitoring: Logs, Alerts
Day 40: LinkedIn Post: "Orchestrating GenAI Workflows with Prefect"
Days 41–50: No-code + Low-code Automation with n8n
Day 41: Install n8n locally (Docker)
Day 42: Create your first Workflow in n8n
Day 43: Automate LinkedIn Posting using n8n + APIs
Day 44: Build an OpenAI-based Email Summarizer in n8n
Day 45: Use Webhooks to trigger LLM Responses
Day 46: Connect Vector DB Search into n8n Workflows
Day 47: Multi-step Approval Flows (Human-in-Loop)
Day 48: n8n + FastAPI integration (Trigger API calls)
Day 49: n8n Error Handling + Retry Strategies
Day 50: LinkedIn Post: "My No-Code GenAI Apps using n8n"
Days 51–60: AutoGen Framework
Day 51: Introduction to AutoGen by Microsoft
Day 52: Install AutoGen, understand Agent concept
Day 53: Build simple AutoGen Assistant (input -> task -> output)
Day 54: Learn how to define UserProxyAgents and AssistantAgents
Day 55: Create a 2-Agent Collaboration (Research -> Summarize)
Day 56: Customize message passing
Day 57: AutoGen Memory Management
Day 58: Build a "Problem Solver" using AutoGen
Day 59: Test concurrent multi-agent coordination
Day 60: LinkedIn Post: "Exploring Microsoft AutoGen for Multi-Agent Systems"
Days 61–70: Multi-Agent Architectures
Day 61: Research: What is A2A (Agent to Agent Communication)?
Day 62: Design: How to pass context from Agent 1 -> Agent 2 -> Agent 3
Day 63: Build a Research Agent + Writing Agent team
Day 64: Add an Editor Agent to review outputs
Day 65: Chain agents using explicit task delegation
Day 66: Error handling between multiple agents
Day 67: Autonomous Planning Agent (Task Prioritization)
Day 68: Dynamic role allocation in Agents
Day 69: Logging and Monitoring Multi-Agent Flows
Day 70: Publish: "Building Multi-Agent AI Systems 🚀"
Days 71–80: Deep Dive into MCP (Model Context Protocol)
Day 71: Understand why MCP was created (problems with Function Calling)
Day 72: Install modelcontextprotocol SDK (Python)
Day 73: Build a simple MCP Server (StdioTransport)
Day 74: Define your Toolset (document reader, searcher, summarizer)
Day 75: Connect LLM to MCP Server
Day 76: Handle Tool Invocations with rich context
Day 77: Extend to multiple tools
Day 78: Create a Workflow Agent using MCP
Day 79: Deploy your MCP server via Docker
Day 80: LinkedIn Post: "Using MCP to Build Smarter Agents!"
Days 81–90: Real Autonomous Projects
Day 81: Design a Full Multi-Agent System: (Researcher -> Writer -> Editor -> Publisher)
Day 82: Use AutoGen Agents with MCP Tools
Day 83: Set up n8n as the Orchestration Layer
Day 84: FastAPI Endpoints to trigger Workflows
Day 85: Vector Search for context retrieval
Day 86: Prefect Flows for retries, failures, notifications
Day 87: Make agents learn from feedback (Self-Improvement)
Day 88: Deploy your full Autonomous System
Day 89: Monitor Logs, Retries, Failures via Prefect and n8n
Day 90: Post: "Building and Deploying an Autonomous Agentic System"
Days 91–100: Grand Projects and Certifications
Day 91: Brainstorm your Final Grand Project idea
Day 92: Project Setup (GitHub Repo + Project Plan)
Day 93: Create Agents (minimum 3 roles)
Day 94: Create APIs, MCP Server, and Connect them
Day 95: Test A2A Messaging + Dynamic Delegation
Day 96: Build a Leaderboard or Scoring Agent
Day 97: Deploy all services with Docker Compose
Day 98: Final Testing and Debugging
Day 99: Polish README, Documentation, Deployment Guide
Day 100: GRAND POST on LinkedIn:
"From Zero to Building Autonomous AI Systems — My 100 Days Journey 🚀"
🌟 This version summary:
1–20
Python + GenAI Core
RAG App
21–40
LangChain + Prefect
Workflow APIs
41–60
n8n + AutoGen
Low-code Automations, Multi-Agents
61–80
MCP + A2A
Build Autonomous Teams
81–100
Full Project
End-to-end System Deployment
Last updated