Language AI courses and concepts

Language AI is the branch of artificial intelligence that works with human language — reading it, writing it, translating it, and converting between speech and text. Language was the hardest thing to hand to a computer, because meaning depends on context that is nowhere in the words themselves. The modern approach sidesteps that problem rather than solving it: turn text into numbers, learn which pieces tend to follow which — which is why these systems became fluent well before they became accurate. Learning the subject means getting comfortable with that conversion first, and then with the particular jobs built on top of it.

The vocabulary of language AI

Two vintages of terminology are mixed together here. One set names the jobs — pulling names out of text, judging sentiment, translating, recognizing speech — and predates the current generation of models by decades. The other describes the machinery underneath, and it is the more useful set to own: token, tokenization and word embedding explain things you will meet directly, from the limit on how much text you can paste in to why usage is counted in pieces rather than words.

Courses on language AI

The question that actually divides these is whether you have a specific language job or a general-purpose model. Pulling fields out of scanned documents, turning recordings into text, giving a search box the ability to answer a plain question — each of those has purpose-built services and courses that teach one of them end to end; working with a general model is a different skill, closer to writing instructions than to configuring a service.

Learn AI — what language models are doing

Use AI — language services and everyday tools

Develop AI — build on and around the models

Language is where most people meet AI first, which makes it a good place to widen out from. Learn AI covers the machine learning underneath, so that a language model's behavior stops being surprising; Use AI is the wider set of things worth doing with these tools in a job that is not about language at all; Develop AI goes into the engineering — the models, the interfaces to them, and the systems that feed them the right context.

← Browse all courses