The Developer Roadmap to AI Engineering · 2026

AI Engineering
Roadmap

The structured path from zero to production-ready AI

Roadmap · guides · projects · career paths — all free for developers

7 Phases~14 months50+ Free resources7 Projects
Open source·No ads·No login required·Free forever

Not just links

Topics, projects & milestones — not a bookmark list.

Built for developers

No math papers. Ship AI products from day one.

Track progress

Check off topics. Saved in your browser forever.

Star on GitHub Feedback
Where are you starting?
01
Phase 1 – AI FoundationsUnderstand the landscape
4–6 weeks·5 topics · 5 resources

Understand how AI/ML works conceptually. No heavy math — just intuition and vocabulary.

  • Click any topic to check it off and track your progress
  • What is AI, ML, Deep Learning, and GenAI — and how they relate
  • Neural networks: inputs, weights, layers, outputs
  • How models learn (gradient descent, loss functions)
  • What LLMs are and how they generate text
  • Key terms: tokens, embeddings, parameters, inference
02
Phase 2 – LLM Setup & ConfigurationGet your environment ready
2–3 weeks·8 topics · 4 resources

Set up your local and cloud AI environment. Know the difference between running, hosting, and calling an LLM.

You have a local LLM running and can call multiple LLM APIs from code.

03
Phase 3 – Prompt Engineering & LLM APIsStart building immediately
3–4 weeks·5 topics · 4 resources

As a dev, you can start building real AI-powered apps right now using APIs — no training needed.

You have a working AI-powered app you built yourself using an LLM API.

04
Phase 4 – RAG & Working with Your Own DataMake AI know your domain
4–5 weeks·5 topics · 3 resources

Learn to feed your own documents, data, and knowledge into AI systems — critical for real-world apps.

You can build a RAG pipeline from scratch and evaluate its quality.

05
Phase 5 – Agentic AIAI that thinks and acts
4–5 weeks·8 topics · 4 resources

Build AI systems that don't just answer — they plan, use tools, and execute multi-step tasks autonomously.

You understand agentic design patterns and have built a working multi-step agent.

06
Phase 6 – Building & Training LLMsGo deep under the hood
6–8 weeks·7 topics · 5 resources

Understand how LLMs are actually built and trained. Know when to fine-tune vs prompt vs RAG.

You can fine-tune an open-source model, understand what happened under the hood, and evaluate the result.

07
Phase 7 – Build & Ship Real ProjectsWhere knowledge becomes mastery
Ongoing·5 topics · 4 resources

Real mastery comes from building. Pick projects that excite you and ship them.

You have shipped 2–3 real AI projects and can discuss AI topics with genuine depth.

Essential Free Tools

Google Colab

Free cloud GPUs for running code

Hugging Face

Models, datasets, free Spaces

Ollama

Run LLMs locally for free

LM Studio

GUI to run local models

LangChain

Build RAG & agent apps

Unsloth

Fast free fine-tuning

Principles for the Journey

Build early, build often: Don't wait until you feel 'ready'. Ship something after Phase 3.
Concepts > memorization: Understand the why. Tools change fast, intuition doesn't.
Your dev skills are a superpower: You can go from idea to working AI app faster than most beginners.
Prompt → RAG → Fine-tune: Always try the simplest approach first before going deeper.
Projects are your portfolio: Even personal projects signal domain expertise better than certificates.

Help make this better

This roadmap is community-driven. If you spot something outdated, missing, or wrong — please say so.

Suggest a missing resourceReport an outdated or broken linkShare your experience after completing a phase
View all discussions on GitHub

Click each phase to expand · Use tabs to navigate sections