Learn AI Agents
Scrimba's Pro course on AI agents, taught by Bob Ziroll across about 2 hours. It covers the step beyond a single LLM call: agents that reason in multiple steps and reach out to functions and APIs to get things done.
Quick answer
Learn AI Agents is Scrimba's intermediate, Pro-tier course on building LLM-powered agents: roughly 2 hours across 30 lessons, taught by Bob Ziroll. You learn how an agent does multi-step reasoning and interacts with its environment by calling functions and APIs, rather than just returning text. It assumes you already know the AI fundamentals and want to make models act, not just answer.
Learn AI Agents
ProTaught by Bob Ziroll (opens in a new tab)
Build LLM agents that reason in multiple steps and call functions and APIs to act on their environment.
View on Scrimba (opens in a new tab)Is it worth your time?
Agents are where a lot of the interesting AI work is heading, and this course gets you past the hand-wavy explanations into how the loop actually works: the model decides, calls a tool, reads the result, and decides again. Bob Ziroll is one of Scrimba's strongest instructors, so the explanations land.
The honest caveat is scope. At two hours this is a focused introduction, not an exhaustive treatment of agent frameworks or production hardening. It also leans on AI fundamentals you are expected to bring with you. If LLM basics are still fuzzy, the agent loop will feel abstract.
What you'll learn
The course centres on the agent loop: how to give an LLM access to functions and APIs, how it decides which to call, and how multi-step reasoning chains together into something that completes a task. You work through this interactively, building the pieces yourself rather than reading about them, which is the right way to understand why agents behave the way they do.
Who it's for, and who should skip it
It fits developers who understand LLM basics and want to build things that take actions, not just generate text. It pairs naturally with the SDK-based agent courses on the catalog.
Skip it if you have not done the AI fundamentals yet. Start with Intro to AI Engineering. Skip it too if you want a specific framework end to end; this teaches the concept, and the project courses go deeper on tooling.
Prerequisites
Comfort with JavaScript and a grasp of how LLM calls work (prompts, tokens, calling a model from code). Intro to AI Engineering is the ideal lead-in.
Where it fits
This sits in the middle of the AI Engineer Path, after the fundamentals. It pairs well with Build Serverless AI Agents with Langbase and Build a Support Agent with Vercel AI SDK, which apply the same ideas with specific tooling.
Free or Pro
This is a Pro course requiring a Scrimba subscription. Pro also covers the full AI Engineer Path, the challenges, the Discord, and certificates. See current plans for pricing in your region. For a free taste of agents first, try Build Serverless AI Agents with Langbase.
Strengths and limits
What it does well: it makes the agent loop concrete, it is taught by a clear instructor, and it is hands-on rather than theoretical.
Where it is limited: it is a short introduction, not a deep dive into frameworks or production concerns, and it assumes AI fundamentals you bring with you.
Related courses and comparisons
- Intro to AI Engineering, the prerequisite fundamentals
- Build Serverless AI Agents with Langbase, a free, applied companion
- Build a Support Agent with Vercel AI SDK, a project-based agent build
- Learn RAG, to ground an agent's answers in your data
No. It is a Scrimba Pro course requiring a subscription. If you want to try agents for free first, Build Serverless AI Agents with Langbase is a free option.
It is strongly recommended. This course assumes you understand LLM basics like prompts, tokens, and calling a model, which Intro to AI Engineering covers.
Bob Ziroll, who also teaches the React courses on Scrimba, so the clear, build-along style carries over.
An LLM that does more than answer: it reasons across multiple steps and calls functions or APIs to act on its environment, then reacts to the results.
This one teaches the concept of agents broadly. The Langbase and Vercel courses apply the same ideas with specific tooling, so they pair well after it.