Prompt file imported from mazleon/llm_learning (.windsurf/workflows/plan.md). Copyright stays with the author.

LLM Learning Project Plan

This document outlines the detailed project plan for learning and implementing various Large Language Model (LLM) frameworks and tools.

📅 Project Overview

• Objective: Gain hands-on experience with modern LLM frameworks and tools
• Duration: 12-16 weeks (flexible)
• Approach: Project-based learning with incremental complexity
• Repository: GitHub Repository

🎯 Learning Objectives

1. Master core LLM concepts and prompt engineering
2. Build applications using LangChain and LangGraph
3. Implement observability with LangSmith and LangFuse
4. Develop multi-agent systems with CrewAI
5. Ensure reliable outputs with Pydantic AI and Agno
6. Complete a capstone project integrating all learned technologies

📚 Learning Modules

Module 1: Foundations & OpenAI SDK (1-2 weeks)

• Topics:
  • LLM fundamentals
  • OpenAI API and SDK
  • Prompt engineering
• Projects:
  • CLI Chatbot
  • Text Summarization Tool

Module 2: LangChain (2-3 weeks)

• Topics:
  • Core components
  • Document processing
  • RAG implementation
• Projects:
  • PDF Research Assistant
  • SQL Database Agent

Module 3: LangGraph (1-2 weeks)

• Topics:
  • Workflow design
  • State management
  • Error handling
• Projects:
  • Enhanced PDF Assistant with LangGraph
  • Multi-Agent Collaboration Workflow

Module 4: Observability (1-2 weeks)

• LangSmith:
  • Tracing and debugging
  • Performance evaluation
• LangFuse:
  • Analytics and monitoring
  • A/B testing

Module 5: CrewAI (2 weeks)

• Topics:
  • Agent design
  • Task orchestration
• Projects:
  • Market Analysis Crew
  • Trip Planning Crew

Module 6: Google ADK (1-2 weeks)

• Topics:
  • Conversational agents
  • Service integration
• Project:
  • Smart Information Kiosk

Module 7: Output Control (1-2 weeks)

• Pydantic AI:
  • Structured outputs
  • Data validation
• Agno:
  • Content moderation
  • Safety guardrails

Module 8: Capstone Project (3-4 weeks)

• AI-Powered Content Creation Assistant
  • Research agent
  • Drafting agent
  • Review agent
  • Integration and deployment

🛠️ Technical Stack

• Languages: Python 3.8+
• Frameworks: LangChain, LangGraph, CrewAI
• Tools: LangSmith, LangFuse, Pydantic AI, Agno
• Version Control: Git/GitHub

📈 Success Metrics

• Completion of all module projects
• Code quality and documentation
• Demonstrated understanding through implementation
• Capstone project completion

📝 Next Steps

1. Set up development environment
2. Begin Module 1: Foundations & OpenAI SDK
3. Complete initial project (CLI Chatbot)

🔄 Updates

• 2025-06-14: Initial plan created

Last updated: 2025-06-14