- Edtech
- ML & Data Science
- Discovery Workshop
- Custom SaaS Development
Enrolment growing, completion flat
An online provider moves every learner through the same fixed sequence, so the ones who already know module three are bored and the ones who did not understand module two are lost. Mastery based pathways fix both without removing instructor oversight.
Representative engagement. This describes a pattern we build rather than one named client: the situation that produces it, how we approach it, and the range of outcomes that kind of work lands in. Figures are stated as ranges or targets, never as a measured result for a specific customer. Our named client work is on the work index.
- Industry
- Edtech
- pathways generated per concept, not per module
- Mastery based
- every routing decision shown with its evidence
- Instructor override
The problem
A fixed curriculum is easy to build and easy to schedule, and it is why completion stalls. Two learners with different starting points get the same next lesson, and one of them is being wasted while the other is being left behind. The instinct is to buy an adaptive engine. The obstacle is that adaptivity is only as good as the assessment behind it, and most course assessment is a quiz at the end of a module, which tells you almost nothing about which specific concept was missed. Instructors are also, reasonably, suspicious. A system that reroutes a learner without explaining why is a system they will override or ignore, and an adaptive platform nobody trusts is a fixed curriculum with extra cost.
What we built
Assessment first. Content is mapped to concepts, and questions are mapped to the concept they test, so a wrong answer identifies a gap rather than producing a score. Pathways are then generated from mastery of concepts rather than from position in a sequence. A learner who demonstrates a concept moves past it, and one who misses it gets targeted practice on that concept rather than a repeat of the whole module. Instructors keep the override and get the reasoning. Every routing decision shows which evidence produced it, which is what makes the difference between a tool teachers use and one they work around. Where an instructor overrides consistently, that is treated as a signal about the model rather than about the instructor.
What changed
Learners stop spending time on material they have already mastered, which is where most of the reclaimed time comes from, and get targeted help where they actually stalled. Published research on mastery based approaches puts the motivation and completion effects in the range below. That is the range this pattern is scoped against rather than a figure we have measured, and the honest test for any particular provider is a single cohort run side by side with the fixed sequence.
Built with
- Python
- TypeScript
- PostgreSQL
- Next.js
- Airflow