Templates

Education

Reference architecture for an AI reading and learning tutor – speech scoring plus curriculum-grounded tutoring with progress tracking, powered by the NAISH Hub.

A learner practises reading aloud or works through lessons on an app or by voice; a teacher watches progress from a dashboard. This template shows a tutor that scores speech, teaches from the curriculum, and tracks each learner.

Education reference architecture

How it works

  1. Channels – learners use an app or voice; teachers use a dashboard.
  2. Gateway & access – handles access and authentication.
  3. AI tutor agent (+ skills) – calls skills: speech scoring for reading fluency, the model for tutoring dialogue, retrieval over the curriculum for grounding, and a progress service to track mastery.
  4. Speech scoring (ASR) – reads recorded audio and writes fluency scores to the progress service.
  5. NAISH Hub – serves the ASR and language models on GPU. Sovereign AI core.
  6. Data & dependencies – a vector store grounds tutoring, a learner-records store tracks progress, and identity gates access.

Where the Hub adds value

  • Local-language literacy – Hub-served ASR and models assess and tutor in the languages learners actually read.
  • Grounded tutoring – retrieval keeps the tutor on the official curriculum.
  • Sovereignty – learner data and models stay on national infrastructure; see Governance & Risk.

Reusable building blocks

LayerThis templateSwap in for your context
ChannelsLearner app / voice + teacher dashboardAny learner or educator client
Agent + skillsSpeech scoring, retrieval, progressSubject-specific skills
Models (Hub)ASR + tutoring LLMCurriculum-tuned models
ServicesProgress serviceWhatever LMS the programme runs
DataLearner records + curriculumAny governed education source