The science, explained in the open.

Knowledge tracing

Knowledge Tracing: How AI Models What a Learner Knows

A stream of right and wrong answers becomes a living estimate of what a learner has mastered.

8 min readRead → Diagnosis

Diagnostic Questions: How Wrong Answers Reveal Mental Models

A well-designed wrong answer doesn't just mark a learner wrong — it names the misconception behind it.

8 min readRead → Trust & regulation

Interpretable AI in Education: Why Explanations Matter

A prediction a teacher can't question is a dead end — and soon, a legal one too.

9 min readRead → Difficulty

The 85% Rule: Why the Best Difficulty Isn't Easy

There's a research-backed sweet spot where learning runs fastest — and it isn't “easy”.

8 min readRead → Root cause

Prerequisite Knowledge Graphs: Finding the Root of a Gap

A learner who looks “weak at math” is often missing one foundation. Here's how to find it.

8 min readRead →