Skip to content

Roadmap.sh Alignment

This page maps the roadmap.sh AI Agents roadmap into this repository's learning-database structure.

Prerequisites

Repository stage: 01 Prerequisites

  • Basic Backend Development
  • Git and Terminal Usage
  • REST API Knowledge
  • Streamed vs Unstreamed Responses

LLM Fundamentals

Repository stage: 02 LLM Fundamentals

  • Transformer Models and LLMs
  • Tokenization
  • Context Windows
  • Token Based Pricing
  • Open Weight Models
  • Closed Weight Models
  • Reasoning vs Standard Models
  • Fine-tuning vs Prompt Engineering
  • Embeddings and Vector Search
  • Understand the Basics of RAG
  • Pricing of Common Models

Generation Controls

Repository stage: 02 LLM Fundamentals

  • Temperature
  • Top-p
  • Frequency Penalty
  • Presence Penalty
  • Stopping Criteria
  • Max Length

Prompt Engineering

Repository stage: 03 Prompt Engineering

  • What is Prompt Engineering
  • Be specific in what you want
  • Provide additional context
  • Use relevant technical terms
  • Use examples in your prompt
  • Iterate and test your prompts
  • Specify length, format, and output constraints

AI Agents 101

Repository stage: 04 Agent Fundamentals

  • What are AI Agents?
  • What are Tools?
  • Agent Loop
  • Perception / User Input
  • Reason and Plan
  • Acting / Tool Invocation
  • Observation and Reflection

Tools and Actions

Repository stage: 05 Tools and Actions

  • Tool Definition
  • Name and Description
  • Input / Output Schema
  • Error Handling
  • Usage Examples
  • Web Search
  • Code Execution / REPL
  • Database Queries
  • API Requests
  • Email / Slack / SMS
  • File System Access

MCP

Repository stage: 06 MCP

  • Model Context Protocol
  • MCP Hosts
  • MCP Client
  • MCP Servers
  • Creating MCP Servers
  • Local Desktop
  • Remote / Cloud

Agent Memory

Repository stage: 07 RAG and Memory

  • What is Agent Memory?
  • Short Term Memory
  • Long Term Memory
  • Within Prompt
  • Vector DB / SQL / Custom
  • Episodic vs Semantic Memory
  • RAG and Vector Databases
  • User Profile Storage
  • Summarization / Compression
  • Forgetting / Aging Strategies

Agent Architectures

Repository stage: 08 Agent Architectures

  • ReAct (Reason + Act)
  • RAG Agent
  • Chain of Thought
  • Planner Executor
  • DAG Agents
  • Tree-of-Thought

Building Agents

Repository stage: 09 Building Agents

  • Manual from scratch
  • Direct LLM API calls
  • Implementing the agent loop
  • Parsing model output
  • Error and rate-limit handling
  • LLM Native Function Calling
  • OpenAI Function Calling
  • Gemini Function Calling
  • Anthropic Tool Use
  • LangChain
  • LlamaIndex
  • Haystack
  • AutoGen
  • CrewAI

Evaluation and Observability

Repository stage: 11 Evaluation and Observability

  • Metrics to Track
  • Unit Testing for Individual Tools
  • Integration Testing for Flows
  • Human in the Loop Evaluation
  • LangSmith
  • Ragas
  • DeepEval
  • Structured logging and tracing
  • Helicone
  • LangFuse
  • OpenLLMetry

Security and Ethics

Repository stage: 12 Security and Ethics

  • Prompt Injection / Jailbreaks
  • Tool Sandboxing / Permissioning
  • Data Privacy + PII Redaction
  • Bias and Toxicity Guardrails
  • Safety + Red Team Testing

Repository Additions Beyond Roadmap.sh

These sections are added to make the site more useful as a professional learning database: