Free · No sign-up · Beginner to interview
A step-by-step roadmap from how LLMs work to RAG, agents and shipping to production, and on to the interview. Every topic is an animated video, a written lesson and a quick quiz.
Beginner 6:44 video
§ 01
Five stages, from your first look at how a language model works to designing AI systems in an interview. Each lesson needs only the ones before it.
Stage 01 · Beginner
What an AI engineer does, how a language model turns text into an answer, how to pick one, and how to talk to it with prompts and code.
3 of 25 lessons live
Stage 02 · Intermediate
Give a model your own documents with embeddings and RAG, then prove your app works with evals.
4 of 19 lessons live
Stage 03 · Advanced
Models that plan, call tools and check their own work: the agent loop, MCP, multi-agent systems and context engineering.
4 of 17 lessons live
Stage 04 · Advanced
Ship it safely and keep it fast, cheap and correct: security, observability, serving, fine-tuning and multimodal.
0 of 22 lessons live
Stage 05 · Interview
Put it all together the way interviews ask: design a real AI system end to end and defend each trade-off.
0 of 6 lessons live
Every topic, in order, with your progress.
See the full roadmap§ 02
Every lesson gives you three ways into the same idea, because understanding sticks when you see it, read it and use it.
01 · Watch
A short animated explainer, narrated, with Vora as your guide: diagrams move, data flows, code runs.
02 · Read
Confident and wrong: the morning after the final
It's the morning after a cricket final, and Arjun's flatmates are arguing about the result. Arjun opens a chat assistant and asks two things. First: "Write a funny message inviting my friends to watch the replay tonight." It writes a brilliant one in seconds.
A written deep dive at your own pace, with code that has been run before it was published.
03 · Practice
Which description of a large language model is most accurate?
A short quiz that explains every answer, right or wrong.
See the whole lesson: What Is a Large Language Model (LLM)? Why It Can Be Confidently Wrong
§ 03
New lessons are added as they pass review.
6:44AI Engineering Beginner 6 min read
A large language model writes a reply by guessing what usually comes next, one piece at a time. That makes it brilliant at writing and explaining, and confidently wrong about facts it never read.
Oct 28, 2026
3:25AI Engineering Beginner 3 min read
An AI model can only see a fixed amount of text at once: its context window. When your text doesn't fit, part of it is cut, and the model seems to forget.
Oct 26, 2026
7:40AI Engineering Beginner 6 min read
Asking the model nicely in the prompt is not a guardrail. A guardrail is a check your code runs around the model, and it can say no whatever the message says.
Oct 25, 2026
4:40AI Engineering Beginner 5 min read
MCP, the Model Context Protocol, is one shared plug between AI apps and tools. Follow Alex asking their assistant to book an hour with Bob, and see 12 custom connections become 7 pieces.
Oct 25, 2026
7:29AI Engineering Beginner 6 min read
An agent is a model inside a loop: it reads the goal and the history, picks the next tool, sees the result, and repeats. Loop engineering is designing that loop on purpose, including how it stops.
Oct 24, 2026
7:04AI Engineering Beginner 6 min read
The model never runs a tool. It writes a call, a tool name plus arguments, and your code checks it, runs it and hands back the result. Here is one round trip, start to finish, at a small cinema.
Oct 24, 2026
§ 04
Developers and students who want to become AI engineers, and engineers preparing for AI engineering interviews. You can start from zero: the roadmap goes one topic at a time, and each lesson builds only on the ones before it.
Yes. Every lesson, video and quiz is free. No paywalls, no pop-ups.
No, there's no sign-up. When you mark a lesson as done, that's saved in your browser only, so it stays on this device and nothing is sent to us.
An animated, narrated video with captions, a written deep dive with callouts and key takeaways, code that is run in a sandbox before it's published, and a short quiz.
Each lesson starts from the author's own notes. Every claim is checked, every code example is run in a sandbox, and the author reviews and signs off on each lesson before it goes live.
There's no fixed schedule: new lessons are added as they pass review. The RSS feed, linked in the footer, announces each one.
Email support@codestoa.com if something isn't working, like a video that won't play or a broken link. Email feedback@codestoa.com to report a mistake in a lesson or suggest a topic. The author reads every message.