AI Foundations
What a model is, how generative models produce output and where hallucination comes from, what makes AI agentic, how enterprises access models, and how to choose one for a task.
- Length
- Short
- Lessons
- 5
- Pages
- 6
- Format
- Self-paced
Ends in a certification once every page and the course check are done.
Before you sell, scope, build or run anything with AI in it, you need a working picture of what the AI does. Five short lessons give you that picture. None of them is about Noxus: everything here holds for any model, from any provider, in any product.
Who this course is for
Everyone who will work with Noxus, in any role. You need no background in AI, statistics or programming. Each technical term gets a one-sentence explanation and a link to the Glossary.
What you will learn
- Tell a predictive model from a generative one, and a task-specific model from a foundation model.
- Say what a decision model returns and where it fits in a process.
- Say what an LLM, an embedding model, a reranker and an OCR model each do inside a process.
- Explain how a generative model produces output, why the same prompt can give different answers, and where hallucination comes from.
- Say what turns a model into an agent, and what an agent needs around it to do real work.
- Recognise what is sold as agentic and is not, and name the mistakes to avoid once an agent is real.
- Name the four layers at which an enterprise holds AI, and say which question each answers.
- Name the routes an enterprise uses to reach models, and the terms that come with them: BYOK, data residency, self-hosting, sovereignty, model independence.
- Choose between a frontier model, a smaller model and no model at all for a given step.
Course contents
- 2.
- 3.What makes AI agentic
One page
- 4.How enterprises access models
One page
- 5.Choosing a model for a task
One page
- 6.Check your understanding
One page
After this course
Platform Foundations shows where these ideas sit in Noxus.