From the intuition behind transformers to training and deploying your own models. Covers the technical side of large language models, diffusion models, and the ecosystem of open and closed foundation models.
Sorted cheapest first. Click a column header to re-sort. Developers and technically curious learners.
| Course | Level | Duration | Price | Certificate | |
|---|---|---|---|---|---|
Nir Diamant (DiamantAI) · GitHub |
intermediate | 40+ runnable notebooks, self-paced | Free | No | Visit |
Google Cloud · Google |
beginner | ~2 hours (4 activities) | Free | Yes | Visit |
Stanford University · YouTube |
advanced | Weekly seminar, Spring quarter (recordings available) | Free | No | Visit |
Microsoft · GitHub |
beginner | 21 lessons, self-paced | Free | No | Visit |
Hugging Face |
intermediate | 12 chapters, self-paced | Free | No | Visit |
Maxime Labonne · GitHub |
intermediate | 3 parts with curated readings and notebooks, self-paced | Free | No | Visit |
Cohere |
beginner | Self-paced modules | Free | No | Visit |
DataTalks.Club · GitHub |
intermediate | 10 weeks | Free | Yes | Visit |
Andrej Karpathy · YouTube |
intermediate | 8 lectures (~15+ hours), ongoing | Free | No | Visit |
Parlance Labs (Hamel Husain and collaborators) · Parlance Labs |
intermediate | 27 talks and workshops | Free | No | Visit |
Scrimba |
beginner | ~11.4 hours | Paid $24/mo |
No | Visit |
DeepLearning.AI and Amazon Web Services · Coursera |
intermediate | ~16 hours (3 weeks) | Free to audit Free to audit · ~$49 |
Yes | Visit |
Udacity |
intermediate | ~56 hours (3 courses, 3 projects) | Paid $249/mo |
Yes | Visit |
No payment required. Several of these are complete university courses.
A large catalogue of runnable notebooks, each isolating one retrieval-augmented generation technique, from basic chunking and query transformations to reranking, hierarchical and graph-based retrieval, and evaluation. It suits developers who have built a basic RAG app and want to systematically try the techniques that improve retrieval quality.
Google Cloud's short beginner path explaining what generative AI and large language models are, how they work, and the principles of responsible AI. A fast, free badge-earning starting point for non-technical professionals before moving to Google's deeper developer paths.
Stanford's seminar series where leading researchers present the latest on transformers, LLMs, world models, state-space models, multimodal systems, and production inference. For practitioners and researchers who want to follow frontier work straight from the people doing it.
Microsoft's 21-lesson course on building generative AI apps, with Learn and Build lessons covering prompting, RAG, embeddings, function calling, agents, fine-tuning, and security in Python and TypeScript. Great for developers who want a free, code-first path from LLM basics to deployable applications.
Hugging Face's hands-on course on LLMs and NLP using the Transformers, Datasets, Tokenizers, and Accelerate libraries, from fundamentals through fine-tuning, demos, and reasoning models. Best for Python developers who want to work with open models directly rather than only via APIs.
A widely starred roadmap in three tracks: optional fundamentals, the LLM Scientist (architecture, pre-training, fine-tuning, preference alignment, evaluation, quantization) and the LLM Engineer (running models, vector stores, RAG, agents, inference optimization, deployment, security). It suits self-directed learners who want a curated map of the best free resources plus runnable notebooks rather than one instructor's course.
Cohere's free learning hub of expert-led modules and guides on how LLMs work, text embeddings, semantic search, prompt engineering, RAG, and tool use, with code labs using Cohere's API. A practical choice for developers who want enterprise-oriented LLM skills explained clearly.
A project-driven community course that goes from agentic RAG with function calling through vector search, orchestration, evaluation and monitoring to an end-to-end assistant and a capstone of your choice. It suits Python developers comfortable with Docker and the command line who want a complete, production-minded LLM application course with optional cohort deadlines and peer review.
Andrej Karpathy builds neural networks from scratch in code, starting with backpropagation and micrograd and ending with a GPT transformer and BPE tokenizer. For programmers comfortable with Python and intro calculus who want to truly understand how language models work under the hood.
A free library of practitioner talks and workshops from Hamel Husain's network, covering LLM evaluation, RAG fundamentals and systematic RAG improvement, building and deploying LLM apps, and a deep track on when and how to fine-tune. It suits engineers who already ship LLM features and want battle-tested advice on evals and reliability from people widely followed on X.
Watch everything free; pay only if you want graded assignments and a credential.
Covers the full LLM lifecycle from transformer architecture and pre-training through instruction fine-tuning, RLHF, and deployment, taught by DeepLearning.AI and AWS engineers. The right course for Python developers with basic ML who want to understand how LLMs are actually built and adapted.
Structured programs with instructors, cohorts, projects or formal credentials.
Scrimba's interactive career path for building AI-powered apps in JavaScript, covering LLM APIs, embeddings and RAG, and agents, with code you edit directly inside the screencasts. It suits web developers who want a guided, hands-on route into AI engineering at a low monthly cost rather than assembling free resources themselves.
Udacity's project-based nanodegree covers generative AI fundamentals, PEFT/LoRA and GRPO fine-tuning, RAG pipelines with ChromaDB, and multimodal apps with CLIP, Whisper, and diffusion models. For working developers who want mentor-reviewed projects and a career-oriented credential.
Build agents, use the Model Context Protocol, and ship with vendor SDKs.
Write prompts that reliably get the results you want from any model.
Text processing, embeddings, sequence models, and modern transformer NLP.