Learn AI with a path that fits.

It is taught by Andrew Ng and focuses on engineering patterns, reflection, tool use and evaluation rather than treating agents as a single prompt.

Andrew Ng keeps the course grounded in business decisions, project selection and realistic AI capabilities instead of code or hype.

It teaches reusable prompting habits rather than a collection of magic phrases, so the lessons transfer across models and tools.

The course assumes no coding background and uses concrete projects, making it a gentle bridge from using AI tools to creating with APIs and data.

It covers current production concerns—tools, knowledge, MCP and orchestration—inside one coherent cloud learning path.

It connects model capabilities to day-to-day work and gives non-technical learners a useful framework for evaluating opportunities and risks.

It combines agent concepts with implementation and community assignments, giving learners a concrete path beyond conceptual demos.

It pairs strong conceptual explanations with the open-source libraries learners are likely to encounter in real model work.

Google combines concise theory with hands-on exercises and has refreshed the material for modern machine-learning topics.

It remains one of the clearest structured starting points for learners who want real machine learning foundations rather than only generative AI tools.

It focuses on the reliability work that separates a demo from an operated AI product: evaluation, versioning, monitoring and tracing.

Its top-down teaching style helps competent programmers create meaningful results early while still developing sound technical intuition.

The breadth makes it a useful orientation for technology professionals who need to understand the AI solution landscape.

The course combines an applied RAG workflow with NVIDIA-hosted lab infrastructure, assessment and certification in a focused one-day format.

Harvard’s CS50 teaching and substantial programming projects give learners a broader computer-science view of AI than an LLM-only course.

A useful starting point if AI terminology feels unfamiliar. The conceptual approach helps you ask better questions about AI claims and applications.

The curriculum connects model theory with practical lifecycle decisions and includes an AWS production perspective.

The short format suits a first step into workplace AI. Activities connect prompting with familiar tasks such as generating ideas and drafting content.

A broader follow-on to introductory training, with activities built around workplace tasks and Google AI tools. Best suited to learners who want to practise across several workflows.

It combines open course material, graded practical work and a no-cost certificate, which is unusual for a technical programme.

It goes beyond prompt-based image creation and teaches the underlying model workflows with a respected open-source library.

The course comes from Hugging Face in partnership with Anthropic and covers MCP from both a user and builder perspective.

It tackles an advanced, practical skill with implementation-oriented material instead of offering another general LLM overview.

Learners write useful model code immediately without spending time configuring a local development environment.

It is brief, current and gives complete beginners the vocabulary needed before committing to a longer technical course.

The operational perspective is useful for people accountable for delivery, not just model creation.

It addresses the failure point many technical courses skip: choosing the right problem and success measure before building a model.

It covers the full system rather than stopping at vector search, including the retrieval and evaluation decisions that determine real-world quality.

It is short, accessible and framed around common work rather than software development.