LTI-ready architecture
Designed around draft platform registration flows, internal identity mapping, and reviewable downstream boundaries.
Stevens IT Review PortalReview BuildSynthetic Examples
Institutional review portal for the Stevens LTI environment
This pre-registration review build presents the architecture, privacy posture, security controls, and synthetic stored embedding records for an LTI-ready AI learning assistant.
Review Section
What this system is
The review environment is structured for institutional inspection and uses concise explanations aligned with the documented architecture.
Review-oriented build
The interface is organized for institutional inspection rather than live course operations.
Privacy-minimized handling
Application-level processing uses pseudonymous learner UUIDs instead of direct identifying information.
Embedding transposition
Learner activity is pseudonymized, converted into typed non-text events, and then transposed into fixed-length embeddings for downstream use.
Review Section
Review readiness
Quick reference for environment status and review-safe operating assumptions.
- Frontend
- Implemented in the review portal
- Backend
- Deferred to a later phase
- Identity model
- Pseudonymous learner UUIDs
- Review content
- Synthetic and non-production
Review Section
What reviewers can inspect
The portal is organized around the principal review topics identified in the docs.
- Data flow from Canvas launch through downstream processing
- Privacy boundaries and excluded data categories
- Security controls and role-separated storage
- Synthetic stored embeddings and metadata for review
Environment readiness
Public environment variables are shown as configured or missing only. Values are not displayed in this review UI.
NEXT_PUBLIC_SUPABASE_URLConfigured
NEXT_PUBLIC_SUPABASE_ANON_KEYConfigured
NEXT_PUBLIC_API_BASE_URLConfigured
NEXT_PUBLIC_APP_ENVConfigured
This review build is intentionally static for this phase. It does not include authentication, backend calls, or live integrations.