Agent Architectures
Common Architectures
RAG Agent
ReAct (Reason + Act)
Short Term Memory
Agent Memory
Building Agents
Manual (from scratch)
Direct LLM API calls Implementing the agent loop Parsing model output Error & Rate-limit handling
Vector DB / SQL / Custom
Within Prompt
Long Term Memory
Episodic vs Semantic Memory
Building Using Frameworks
Tree-of-Thought
CrewAI
AI & Data Scientist
AutoGen
Smol Depot
LlamaIndex
Chain of Thought (CoT)
Planner Executor
Other Architecture Patterns
DAG Agents
LLM Native "Function Calling"
OpenAI Functions Calling
OpenAI Assistant API
Gemini Function Calling
Anthropic Tool Use
Example Usecases
Personal assistant
Agent Loop
Code generation
Data analysis
Web Scraping / Crawling
Tool Definition
Name and Description Input / Output Schema Error Handling Usage Examples
Tools / Actions
Prompt Engineering
NPC / Game AI
Model Context Protocol (MCP)
MCP Client
MCP Hosts
Core Components
MCP Servers
Code Execution / REPL
Database Queries
Web Search
Examples of Tools
API Requests
Email / Slack / SMS
File System Access
What is Agent Memory?
Deployment Modes
Local Desktop
Remote / Cloud
Be specific in what you want
Writing Good Prompts
What is Prompt Engineering
Iterate and Test your Prompts
Specify Length, format etc
Prompt Engineering Roadmap
Creating MCP Servers
Evaluation and Testing
Understand the Basics of RAG
Embeddings and Vector Search
Pricing of Common Models
AI Agents 101
What are AI Agents?
What are Tools?
Provide additional context
Use relevant technical terms
Use Examples in your Prompt
Forgetting / Aging Strategies
Summarization / Compression
User Profile Storage
RAG and Vector Databases
Maintaining Memory
Metrics to Track
Unit Testing for Individual Tools
Integration Testing for Flows
Human in the Loop Evaluation
Frameworks
Observability Tools
Debugging and Monitoring
Structured logging & tracing
LangSmith
DeepEval
Ragas
LangSmith
LangFuse
Helicone
openllmetry
1 Perception / User Input
Streamed vs Unstreamed Responses
Reasoning vs Standard Models
Fine-tuning vs Prompt Engineering
✓ AI Engineer Roadmap
✓ AI and Data Scientist Roadmap
✓ MLOps Roadmap
✓ AI Red Teaming Roadmap
✓ Prompt Engineering Roadmap
Temperature
2 Reason and Plan
3 Acting / Tool Invocation
4 Observation & Reflection
Presence Penalty
Stopping Criteria
Frequency Penalty
Max Length
Top-p
Generation Controls
Understand the Basics
Model Families and Licences
Open Weight Models
Closed Weight Models
Backend Beginner Roadmap
Model Mechanics
Transformer Models and LLMs
Context Windows
Token Based Pricing
Tokenization
Data Privacy + PII Redaction
Basic Backend Development
REST API Knowledge
Git and Terminal Usage
LLM Fundamentals
AI Agents
Learn the Pre-requisites
Git and GitHub Roadmap
API Design Roadmap
Find the detailed version of this roadmap along with other similar roadmaps
roadmap.sh
Langchain
Haystack
Visit the following relevant tracks
AI Engineer
Bias & Toxicity Guardrails
Safety + Red Team Testing
All Roadmaps →
Prompt Injection / Jailbreaks
Tool sandboxing / Permissioning
Security & Ethics