In production

The learning manifold.

Every outcome is a location. Difficulty is curvature. Each learner is a position that moves. In production in schools across India.

The premise

Learning has a shape.

Most systems treat skills as a flat list: a learner has them or doesn't. Learning has structure. Concepts connect, skills depend on other skills, and a learner has a position in that structure. Activities reveal where they are. Time moves them through it.

The axes

Every subject has its own dimensions.

English isn't measured the same way as Maths. Maths isn't measured the same way as Science. The shape of what you're learning determines the shape of how it's measured. Each subject has its own coordinate system: the same structural framework, different axes.

Subject

English

  • Accuracy

    getting to the right answer

  • Pronunciation

    clear sound production

  • Fluency

    smooth, well-paced delivery

  • Vocabulary

    range and choice of words

  • Content

    relevance and substance of ideas

Calibrated against CEFR — the Common European Framework of Reference.

Subject

Maths

  • Accuracy

    getting to the right answer

  • Procedural Fluency

    efficient, flexible execution

  • Conceptual Understanding

    grasping why, not just how

  • Reasoning

    justifying, generalising, pattern-spotting

  • Representation

    moving between symbolic, visual, concrete forms

Calibrated against the TIMSS framework — content × cognitive domains.

Subject

Science

  • Accuracy

    correct facts and applications

  • Inquiry

    predicting, hypothesising, explaining

  • Conceptual Understanding

    grasping underlying models

  • Representation

    diagrams, graphs, equations

  • Vocabulary

    domain-specific scientific language

Calibrated against the TIMSS Science Framework and NGSS practices.

Notice the universal axes. Accuracy appears in every subject: the universal “did you get to the right answer?” axis. Conceptual Understanding and Representation recur across Maths and Science. Each subject also has its own irreducible dimensions: Pronunciation belongs to English, Procedural Fluency to Maths, Inquiry to Science. The shape changes; the geometry does not.

The English and Maths geometries are built; Science has its dimension system defined, and its geometry is next.

The principle

Honest measurement, not invented numbers.

When an activity doesn't measure a dimension, we don't fake a score. We say nothing. A listening activity measures Accuracy, and only Accuracy. A speaking activity measures all five English dimensions. A reading-aloud activity measures Accuracy, Pronunciation and Fluency, but not Vocabulary or Content, because the child didn't choose the words.

A learner's position fills in as they cover more activity types. We never invent a score to fill a gap. That keeps the framework honest, maps cleanly to schools that don't run the full activity suite, and gives a natural progress story.

The calibration

Anchored to international standards.

The coordinate system is ours. The calibration anchors are public, international and standards-based.

For English, we calibrate against CEFR, the Council of Europe's six-level framework, used by language assessment bodies in over forty countries. CEFR's can-do descriptors define what a learner at A1, A2, B1, B2, C1 or C2 can actually do with the language. Our dimensions are the measurement axes; CEFR is where points on those axes correspond to recognisable proficiency levels.

For Maths, we calibrate against TIMSS, the IEA's international comparative study, which defines mathematical proficiency across content and cognitive domains. For Science, we draw on the TIMSS Science Framework and NGSS, both publicly published frameworks defining the practices and core ideas of scientific learning.

CEFRTIMSSNGSS

The cost of learning

Why ability gaps widen.

C = ρλ²

Cognitive cost equals complexity density times the square of how fast a given learner propagates understanding. Small differences in λ become large differences in cost exactly where the space is densest, which is why ability gaps widen at higher abstraction.

Curricula

One space, many charts.

A curriculum doesn't define the knowledge; it lays a set of coordinates over it. A child learning to add fractions in Cambridge Primary and a child learning the same thing in an Indian state-board classroom occupy the same region of the space. Adding a curriculum doesn't mean starting again. It means a new chart on a geometry that's already there.

Mapped so far

  • Cambridge Primary English
  • Cambridge Primary & Lower Secondary Mathematics
  • CEFR
  • UK National Curriculum
  • NCERT / CBSE
  • Tamil Nadu State Board
  • Telangana State Board
  • Montessori (ages 3–8)

Ages 3–8

Where it starts.

The early years have their own geometry. We built it from the Montessori method's own ordering of materials: how abstract the work is, how hard, how much of the hand and the senses it uses, and when it's left to the child. No “ahead”, no “behind”; every child at their own pace.

Maths

Solver first.

In maths, every question and every answer is computed by a solver before any language is generated. The language model explains; it never decides what's correct.

The trajectory

Learning is a path, not a level.

Most assessment systems give a learner a level. You're at A1. You're at Year 3 expected. A level is a snapshot. It tells you where the learner is, but not where they came from or where they're going.

We measure a learner's position continuously. That position is a vector, and it changes over time. The change is the learning. A teacher can see not just where a child is today, but where they were last week, which dimensions are growing fastest, and where the next step should take them.

It's in production. Our client Chrysalis uses it across its schools in India, and Outloud English, our own product, is built on it.

For researchers

We work with researchers.

The coordinate system produces an unusual kind of dataset: continuous, dimensional, cross-subject, and anchored to international standards. We work with academic researchers on learning trajectories, calibration and adaptive sequencing.