Fundamentals of Large Language Models

Understand the models behind chatbots, practical prompting and multimodal AI.

Curriculum

LLMs, transformers, attention and text generation.
Prompting, generative AI, work and human creativity.
Multimodal AI and embeddings across different data types.

Before you start

No coding background required; provided code can be run and interpreted.

Who it suits

Our assessment, based on the syllabus.

Product or operations manager

Understand why model outputs vary and how to judge them.

Knowledge worker

Use prompting more deliberately.

Founder evaluating AI products

Discuss model capabilities and limitations credibly.

Expected outcomes

Explain the basic mechanisms behind generative AI tools.
Practise prompting and interpreting model outputs.
Recognise model limitations and ethical issues.

Example application

Explain why an assistant needs source checking before staff rely on its answers.

Limits

Foundational rather than advanced agent implementation.
Not a coding qualification or an academic-credit course.

How current is it?

Part of Universal AI, launched in May 2026. Covers LLMs and multimodality; the emphasis is on underlying concepts.

When to choose it

Consider this for an MIT introduction to how LLMs work and how to interpret their outputs. It suits people who can spend several weeks on the concepts before choosing tools.

Sources

Official course details ↗learn.mit.edu
2026 program launch ↗news.mit.edu

Checked 20 October 2026. Confirm the current offer on the provider’s page before paying.