🗓️ 21032026 2100
A structured reading and learning plan progressing from LLM fundamentals to building custom AI tool implementations
Phase 1: Foundations
Goal: understand how LLMs work and how they interact with tools
| Resource | Type | Link |
|---|---|---|
| Andrej Karpathy — "Intro to Large Language Models" | Video | YouTube |
| Anthropic — Tool Use Documentation | Docs | docs.anthropic.com |
| OpenAI — Function Calling Guide | Docs | platform.openai.com |
Zettelkasten: llm_tool_use
Phase 2: Agent Patterns
Goal: learn how agents are architected and when to use which pattern
| Resource | Type | Link |
|---|---|---|
| Lilian Weng — "LLM Powered Autonomous Agents" | Blog | lilianweng.github.io |
| Andrew Ng — Agentic Design Patterns (DeepLearning.AI) | Video | deeplearning.ai |
| Anthropic — "Building Effective Agents" | Blog | anthropic.com |
Zettelkasten: agentic_design_patterns, agent_memory_and_state
Phase 3: MCP & Tool Infrastructure
Goal: understand the protocol layer that connects agents to tools at scale
| Resource | Type | Link |
|---|---|---|
| MCP Official Documentation | Docs | modelcontextprotocol.io |
| MCP Quickstart — Building a Server | Tutorial | modelcontextprotocol.io |
| MCP Quickstart — Building a Client | Tutorial | modelcontextprotocol.io |
Zettelkasten: model_context_protocol, custom_mcp_servers
Phase 4: Build Your Own
Goal: hands-on implementation to solidify understanding
| Project | Description | Starting Point |
|---|---|---|
| Custom MCP Server | Expose an internal API or database via MCP. Start with stdio transport, then add HTTP | MCP Server Quickstart |
| Agent with Tool Use | Build an agent that uses tools to accomplish a multi-step task | Anthropic SDK, LangChain Agents, LlamaIndex Agents |
| Multi-Agent System | Combine multiple agents using your custom MCP tools — try coordinator or agent-as-tool pattern | coordinator_router_pattern, agent_as_tool_pattern |
Phase 5: Advanced Topics
Goal: deeper exploration for production-grade systems
| Topic | What to Explore |
|---|---|
| RAG (Retrieval-Augmented Generation) | Embedding models, vector databases, chunking strategies, hybrid search |
| Agent Evaluation & Observability | Tracing agent runs, measuring tool call accuracy, detecting loops/hallucinations |
| Production Concerns | Guardrails, cost management, latency optimization, rate limiting, error recovery |