AI & Data · Beginner → Advanced

AI Engineer Roadmap

LLMs → prompting → embeddings → RAG → tools → agents → evaluation → security.

Your progress

0/9 nodes · 122h estimated

Skills

Projects

Phases

Phase 1

LLM foundations

Understand modern AI systems.

ML fundamentals

Training, inference, embeddings, loss and evaluation.

Subtopics

• ML fundamentals fundamentals
• ML fundamentals core concepts
• ML fundamentals practical patterns
• ML fundamentals production considerations

Practice

• Build a small exercise for ML fundamentals
• Debug or test the implementation
• Explain the trade-offs in your own words

Checkpoint

You can explain the core concepts, implement them without a tutorial, and debug a small real-world example.

~12hOutcome: Understand model behavior.

LLM concepts

Tokens, context, sampling, reasoning and multimodality.

Subtopics

• LLM concepts fundamentals
• LLM concepts core concepts
• LLM concepts practical patterns
• LLM concepts production considerations

Practice

• Build a small exercise for LLM concepts
• Debug or test the implementation
• Explain the trade-offs in your own words

Checkpoint

You can explain the core concepts, implement them without a tutorial, and debug a small real-world example.

~12hOutcome: Choose models intelligently.

Model APIs

Structured output, streaming, retries, rate limits and cost.

Subtopics

• Model APIs fundamentals
• Model APIs core concepts
• Model APIs practical patterns
• Model APIs production considerations

Practice

• Build a small exercise for Model APIs
• Debug or test the implementation
• Explain the trade-offs in your own words

Checkpoint

You can explain the core concepts, implement them without a tutorial, and debug a small real-world example.

~10hOutcome: Integrate models safely.

Phase 2

AI application patterns

Turn models into useful product features.

Prompt engineering

Instructions, examples, constraints and structured prompts.

Subtopics

• Prompt engineering fundamentals
• Prompt engineering core concepts
• Prompt engineering practical patterns
• Prompt engineering production considerations

Practice

• Build a small exercise for Prompt engineering
• Debug or test the implementation
• Explain the trade-offs in your own words

Checkpoint

You can explain the core concepts, implement them without a tutorial, and debug a small real-world example.

~8hOutcome: Create predictable behavior.

Embeddings & vector search

Similarity, chunking, metadata and retrieval.

Subtopics

• Embeddings & vector search fundamentals
• Embeddings & vector search core concepts
• Embeddings & vector search practical patterns
• Embeddings & vector search production considerations

Practice

• Build a small exercise for Embeddings & vector search
• Debug or test the implementation
• Explain the trade-offs in your own words

Checkpoint

You can explain the core concepts, implement them without a tutorial, and debug a small real-world example.

~10hOutcome: Build semantic search.

RAG

Ingestion, retrieval, reranking, citations and freshness.

Subtopics

• RAG fundamentals
• RAG core concepts
• RAG practical patterns
• RAG production considerations

Practice

• Build a small exercise for RAG
• Debug or test the implementation
• Explain the trade-offs in your own words

Checkpoint

You can explain the core concepts, implement them without a tutorial, and debug a small real-world example.

~16hOutcome: Build grounded assistants.

Phase 3

Agents & production

Build, evaluate and secure systems that act.

Tool calling

Schemas, validation, permissions and execution.

Subtopics

• Tool calling fundamentals
• Tool calling core concepts
• Tool calling practical patterns
• Tool calling production considerations

Practice

• Build a small exercise for Tool calling
• Debug or test the implementation
• Explain the trade-offs in your own words

Checkpoint

You can explain the core concepts, implement them without a tutorial, and debug a small real-world example.

~10hOutcome: Connect models to tools.

AI evaluation

Datasets, graders, traces and regression tests.

Subtopics

• AI evaluation fundamentals
• AI evaluation core concepts
• AI evaluation practical patterns
• AI evaluation production considerations

Practice

• Build a small exercise for AI evaluation
• Debug or test the implementation
• Explain the trade-offs in your own words

Checkpoint

You can explain the core concepts, implement them without a tutorial, and debug a small real-world example.

~14hOutcome: Measure AI quality.

Production AI app

Project

Ship RAG, tools, evaluation and observability.

Subtopics

• Production AI app fundamentals
• Production AI app core concepts
• Production AI app practical patterns
• Production AI app production considerations

Practice

• Build a small exercise for Production AI app
• Debug or test the implementation
• Explain the trade-offs in your own words

Checkpoint

You can explain the core concepts, implement them without a tutorial, and debug a small real-world example.

~30hOutcome: Build a portfolio AI system.