AI moves weekly—your learning should too
After 5 years of building and sharing educational content for free, we launched our first-ever course
Seven months ago, after 5 years of building and sharing educational content for free, we launched our first-ever course: From Beginner to LLM Developer — a full-stack, no-nonsense course to help engineers build real-world LLM applications from scratch.
At the time, two things were already clear:
The age of LLM development as a distinct, essential software discipline had dawned.
This era will keep evolving pretty fast, and so will learning curricula, contrary to how I learned AI in school.
We are all about these two points, which can be summarized as "equipping every developer with the skills to harness the most powerful, readily available AI tools (LLMs, RAG techniques, fine-tuning and more)".
So...
Today, we are thrilled to announce a significant course content update, reinforcing our commitment to keeping you at the forefront of this rapidly evolving field!
What’s Updated?
We push small updates here and there almost every day. Still, since October 2024, the field has evolved A LOT: new models, reasoning models(!), techniques, MCP, A2A, RFT and more.
If you already have our course (or if you get it), our promise to you is to ensure the course “From Beginner to LLM Developer” gives you permanent access to state-of-the-art, up-to-date resources.
This isn’t just a course; it’s a living curriculum designed to grow with the technology, ensuring your skills are always current.
Here are just some highlights of the changes we've made and lessons we've added:
Run LLMs Your Way with Ollama: Our new “Ollama Tutorial: Running Deepseek Distill Locally” lesson teaches you to set up and manage powerful open-source models on your own local machine. This is invaluable for projects demanding data privacy, offline access, or cost-effective experimentation with various architectures.
Master Next-Generation Reasoning Models: Explore “An Introduction to Reasoning Language Models” (like OpenAI’s o-series, DeepSeek’s r-series, Google’s Gemini Pro 2.5). Learn how these models go beyond next-token prediction to truly plan, think, and solve complex problems – a cornerstone for building sophisticated agents and advanced RAG systems.
Achieve Expert-Level Performance with Reinforcement Fine-Tuning (RFT): Our lesson on “Reinforcement Fine-Tuning for OpenAI” uncovers how to use RFT with models like o4-mini to create “expert models” highly specialized for your specific tasks, often with surprisingly little data. This is key for applications demanding top-tier accuracy in niche domains.
Integrate Visuals: LLM Native Image Generation & Editing with OpenAI’s new image API: Go multimodal! “Generating and Editing Images via the OpenAI API” shows you how to programmatically create and modify images directly using LLMs like GPT-4o. A crucial skill for building richer, more engaging AI applications.
Build for Interoperability with Anthropic’s Model Context Protocol (MCP): Our new lesson “Extending LLM Capabilities with Anthropic’s Model Context Protocol (MCP)” helps you get to grips with this open standard designed to simplify how AI models connect with external tools. Essential knowledge for creating scalable, future-proof agentic systems that aren’t locked into a single ecosystem.
Plus:
📹 More Video Content: We’ve converted even more text lessons into an engaging video format to cater to different learning styles.
✨ General Enhancements: You’ll find updates and polish throughout the course to ensure clarity, incorporate new model releases, and reflect the latest best practices in LLM development.
If you don't have the course yet, the first 10 lessons — including your first real code notebook — are now free. No card required or anything.
If you decide to go with the full course, we also offer a 30-day refund guarantee :)
This update also kicks off our live June cohort, with an AMA with our CEO Louie Peters on Sunday, June 1st.
In a field evolving this quickly, timing is leverage. You don’t want to master prompt tuning when RFT + tool use becomes the new minimum. This cohort helps you keep pace — and stay confident — as the field shifts.
For Newcomers: From Beginner to Advanced LLM Developer is Still the Most Complete Path to LLM Mastery
While we’ve integrated the latest breakthroughs, the foundation of this course remains the same: to arm you with a timeless LLM development toolkit. The fundamental principles of Prompt Engineering, Data Curation, Retrieval-Augmented Generation (RAG), Fine-Tuning, Tool Use, and Agentic Design are skills that will evolve with the models, not become obsolete. This course teaches you how to learn and adapt as the models themselves change, ensuring your skills remain valuable. As we stated in our original thesis, technical development on top of foundation LLMs is here to stay and will always provide better results for specific tasks in specific industries relative to out-of-the-box foundation models and no-code customization.
This is still a full-stack, 90+ lesson course with code notebooks, project templates, and industry-tested techniques. You’ll go from ideation and choosing a suitable LLM application to collecting data, iterating on many advanced techniques, integrating industry expertise, and deployment with a certifiable RAG project. You’ll work with Python, OpenAI, Llama 3, Gemini, LlamaIndex, LangChain, Hugging Face, Gradio, and many other amazing tools. We remain unaffiliated, covering the best options in the ecosystem so you can make informed decisions.
And it’s not just technical. We also cover the non-technical skills critical for launching successful LLM products — things like integrating domain expertise, evaluating model behavior, and designing for trust and usability.
The only prerequisite is some Python knowledge (or basic programming). You’ll still build a working, certifiable RAG project, receive instructor support in our dedicated Discord channel, and gain the critical non-technical skills essential for building great LLM products.
If you believe, like we do, that technical development on top of foundation models will drive the next phase of adoption, this is your way in.
But don’t just take our word for it — engineers across industries have used this course to launch products, land roles, and upskill fast. Check the reviews below to see how it’s working in the wild.
“Best course out there to become an AI engineer. Planning to build my own startup based on the learnings.” — Abhijit L
“Beyond covering fundamental LLM concepts, the course delves into critical decision-making—understanding when and how to apply these technologies effectively.” — Mario Giraldo (AI Consultant)
“Well-structured and approachable…makes the learning process incredibly smooth and accessible.” — Dan Duggan (AI Sales Professional)
“From zero to hero as an LLM Developer…a clear path to build practical applications.” — Luca Tanieli (Software Developer)
Who This Is For (Hint: Probably You)
If you have some developer skills and know a bit about Python. Yep, that's it!
Otherwise, if that is not you, we have this offer to teach you Python specifically for working with LLMs as a new way to learn programming. Check it out here.
👉 Start building — free access to 10 lessons now
p.s. If you’re leading AI efforts at a company or building something ambitious, we offer private team versions of the course and affiliate partnerships. Let’s talk: louis@towardsai.net