AI Engineer Roadmap

1. Introduction

What is an AI Engineer? Roles and Responsibilities Impact on Product Development AI Engineer vs ML Engineer

2. Working With LLMs

How LLMs Work

Core LLM Elements

Tokens Context Sampling Parameters Temperature Top-K Top-P Repetition Penalties

Common Terminology

AI vs AGI Large Language Model (LLM) Embeddings Training Inference Vector DBs AI Agents RAG Fine-tuning

3. Prompt vs Context Engineering

Prompt Engineering

Zero-Shot Few-Shot ReAct CoT Input Format Function Calling Prompt Caching Streaming Responses System Prompting Role & Behavior Context & Constraints Structured Output

Context Engineering

External Memory RAG and Dynamic Filters Context Compaction Context Isolation

4. AI Models

Type of Models

Pre-trained Models Closed vs Open Source Self-Hosted Models

Choosing the Right Model

Closed Models (Claude, Gemini, OpenAI) Open Source (Llama, DeepSeek, Qwen)

Hugging Face

Hugging Face Tasks Hugging Face Hub Transformers.js

Local & Platforms

Ollama LM Studio OpenRouter

APIs & SDKs

OpenAI Response API Claude Messages API Google Gemini API Hugging Face Inference SDK

5. Embeddings & Vector Databases

What are Embeddings

Semantic Search Data Classification Recommendation Systems Anomaly Detection

Embedding Models

Proprietary (OpenAI, Gemini, Cohere) Open Source (Sentence Transformers, Jina)

6. Vector Databases

Popular Vector DBs

Chroma Pinecone Weaviate FAISS Qdrant Supabase (pgvector)

Implementing Vector Search

Indexing Embeddings Performing Similarity Search

7. Retrieval Augmented Generation (RAG)

What are RAGs?

RAG Usecases RAG vs Fine-tuning

Implementing RAG

Chunking Embedding Vector Database Retrieval Process Generation

Ways of Implementing

Using SDKs Directly Langchain Llama Index Haystack

8. AI Agents

Concepts

Agents Usecases ReAct Prompting Tools & Function Calling Multi-agents

Building AI Agents

Custom Agents (from scratch) CrewAI LangGraph AutoGen OpenAI Swarm

Cloud & Frameworks

OpenAI AgentKit Claude Agent SDK Vertex AI Agent Builder

9. Model Context Protocol (MCP)

Core Components

MCP Host & Client MCP Server Data & Transport Layer

Developing with MCP

Building an MCP Server Building an MCP Client Connecting (Local & Remote)

10. AI Safety and Ethics

Safety Issues

Prompt Injection Attacks Security and Privacy Bias and Fairness

Best Practices

Content Moderation APIs Adversarial Testing Robust Prompting Constraining Outputs

11. Evaluation & Observability

Observability Tools

Tracing & Logging Cost / Latency Monitoring Production Monitoring

Evaluation (Evals)

LLM-as-a-judge Evaluation Metrics Prompt Versioning

12. Multimodal AI

Multimodal Concepts

Multimodal AI Multimodal AI Use Cases Image Understanding Image Generation Video Understanding Audio Processing Speech-to-Text Text-to-Speech

APIs and Frameworks

OpenAI Vision API DALL·E API Nano Banana API Whisper API LangChain for Multimodal Apps LlamaIndex for Multimodal Apps

13. Development Tools

AI Code Editors

Development Tools Cursor Windsurf Claude Code Codex Replit