Skip to main content
EDU-01 · Education

Personalized Learning Tutor

Outcome​

Give each learner a tutor that grounds every explanation, hint and practice item in the district-approved curriculum and the learner's own mastery evidence, adapting pace and representation without ever revealing assessment answers or acting as a grade of record.

Portfolio attributeValue
Primary personaStudent
Primary aggregateLearnerMasteryProfile
Cosmos DB containers2
Grounding datacurriculum, content, assessments, learner profile
Development estimateUSD 60-220 monthly dev estimate
Runtime optionsPython FastAPI and C# ASP.NET Core

Productivity hypothesis​

These measures reflect the workflow, evidence, decision boundary, and industry risk for this use case.

Impact typeMeasureConservative internal assumptionCustomer measurement
ProductivityTutoring task cycle time10-18% reduction(baseline median tutoring-task minutes - assisted median minutes) / baseline median minutes * 100
Quality or riskRework rate3-8% reduction(baseline corrected or repeated response rate - assisted rate) / baseline rate * 100

Important: Conservative internal pilot assumptions only; not public benchmarks, commitments, or expected customer results. Validate against a comparable baseline or matched control cohort with mandatory human review.

Rationale: Curriculum, assessments, approved content, and learner profiles support faster tailored explanation preparation.

Pilot guardrail: Educators retain control over instructional suitability and all student-impacting decisions.

Customer validation plan​

  1. Baseline the current workflow with representative users and tasks.
  2. Run the accelerator on the same task mix with human review retained.
  3. Measure tutoring task cycle time and rework rate with the documented formulas.
  4. Replace internal assumptions with customer-observed results before a business case.

Why Azure Cosmos DB​

This AI solution needs an operational data plane, not only a vector index:

  • Hybrid vector and full-text retrieval combines semantic recall with exact identifiers, codes, names, and domain language.
  • Hierarchical partition keys and metadata predicates keep tenant and domain boundaries in the normal query path.
  • Operational JSON, source chunks, workflow and agent state, feedback, and audit records remain together for low-latency reads and governed writes.
  • Change feed reacts to source and operational updates without polling.
  • Autoscale throughput, configurable consistency, and multi-region distribution support variable demand and enterprise availability.

Domain workflow​

Domain workflow

StageActorInput → outputApproval
1. Mastery diagnosis from attempt historyMastery estimatorPracticeAttempt → MasteryStatePolicy controlled
2. Standard-aligned content retrievalRetrieval orchestratorMasteryState → CurriculumChunkPolicy controlled
3. Scaffolded explanation and hintingPersonalized learning tutorCurriculumChunk → TutoringTurnPolicy controlled
4. Adaptive practice and misconception taggingStudent with tutorTutoringTurn → PracticeAttemptPolicy controlled
5. Teacher review of proposed learning pathTeacher of recordLearningPathPlan → approvedLearningPathHuman required

Technical architecture​

Technical architecture

Cosmos DB data model​

ContainerHierarchical partition keyDocumentsSearch paths
learnerMasteryProfiles/tenantId, /learnerIdLearnerProfile, MasteryState, PracticeAttempt, TutoringTurn, LearningPathPlanOperational fields only
curriculumLibrary/tenantId, /standardCodeCurriculumChunk/chunkVectorEmbedding, /chunkVector

Query contract​

QueryContainerScopePurpose
MasteryGapsForCourselearnerMasteryProfilessingle-partitionIdentify standards below the mastery threshold in a course to drive the next practice set.
LevelAppropriateContentRetrievalcurriculumLibrarysingle-partitionRetrieve approved explanations for a standard matched to the learner's reading band and preferred representation.
MisconceptionTrendlearnerMasteryProfilessingle-partitionSurface repeated misconception tags on a standard so the tutor changes representation instead of repeating the same hint.
SessionTranscriptForTeacherlearnerMasteryProfilessingle-partitionReturn a cited tutoring transcript for teacher oversight and academic-integrity review.

Python and C# services route each document type to its declared container and hierarchical partition key. ETag-based writes protect mutable aggregates, while append-only records and bounded cross-partition queries preserve scale.

Azure deployment​

Azure deployment

Clone the source repository, authenticate with Azure, and run:

Set-Location accelerators/education/personalized-learning-tutor
az login
./Deploy.ps1 -Runtime python -Location eastus2

Change python to csharp to deploy the .NET implementation.

Explore the implementation​

Return to all accelerators