EMI Product Alignment: What We Already Have
Principle: Extend, Don’t Rebuild
Nearly every EMI requirement maps to something Learneris already has built or partially built. The strategy is to configure, extend, and rebrand existing systems — not build from scratch.
Mapping: EMI Policy Needs → Existing Learneris Capabilities
1. “AI integration in English teaching” (Step 3 of HCMC De An)
What we have:
- ai-gateway — Multi-model AI router (GPT-4o, Claude Sonnet 4, Gemini 2.5 Flash, image gen via FLUX/Imagen). Already handles chat, image generation, credit billing, failover. All Learneris products route through this.
- AI-Course-fresh — 4 lesson-scoped AI chatbot personas already live for students (natural/artificial classifier, creative art describer, deliberate-error AI for critical thinking, safety advisor). Uses Azure OpenAI with child-safe system prompts.
- LMS — AI content generation (Claude Sonnet 4.6 via SSE streaming) creates lesson HTML and MCQ quizzes from curriculum standards. AI-assisted grading with per-criterion rubric scoring.
- creator.learnery — Full AI course creation pipeline: title → objectives → outline → slides → PPTX export. 7+ output languages including English and Vietnamese.
What to extend for EMI:
- Add new AI chatbot personas to AI-Course-fresh for English practice (conversation partner, vocabulary drill, pronunciation scenarios)
- Add
language parameter to lesson-agent prompts (currently English-only output — add bilingual/CLIL output mode)
- Add EMI-specific content types in creator.learnery (CLIL lesson template, bilingual slide template)
- ai-gateway: no changes needed — it’s already model-agnostic and content-type-routed
Effort: Small-Medium — new prompts and templates, not new infrastructure.
What we have:
- LMS (lms.learneris.com) — Full teacher+student platform with:
- Course authoring (6 block types: text, file, HTML, AI content, AI App, embed)
- Assignment management with rubrics, scoring, time limits, max attempts
- AI quiz generation (3-15 MCQ from curriculum KB)
- Analytics dashboard (completion rates, grade distribution, weekly engagement, top performers)
- Student dashboard (enrolled courses, assignments, calendar, grades)
- Multi-org / role switching (teacher at school A, student at school B)
- Full EN/VI bilingual UI (i18next with runtime toggle)
- 105+ Bubble data types with Zod schemas
- Class management, student enrollment, org management
What to extend for EMI:
- Create EMI-specific course templates pre-loaded with CLIL structure
- Add “English exposure hours” tracking field to student progress (simple schema addition)
- Add school-wide admin view (currently teacher-scoped analytics — extend to org-level rollup)
- Surface the existing
report, reportbatch, recommendation schemas as actual pages (they’re defined but have no UI yet)
Effort: Small — the platform is ready; it needs EMI-specific content and a few dashboard views.
3. “Bilingual digital content repositories” (Step 3 of HCMC De An)
What we have:
- AI-Course-fresh — Already has:
- Full bilingual translation file (
en.json) covering all TH0 lessons
- VN/EN language toggle built into student and teacher HTML books (saves to localStorage)
- EMI school variant in PD training data — every activity has 3
schoolVariants: "emi", "vn-elite", "normal". EMI variant uses English terminology and global examples.
- HCAI Engine bilingual mode — JSON-driven slide engine with
"bilingual": true flag per lesson, biText(vn, en) helper for dual-language rendering
- 5 grades × 16 lessons × 35 periods = 80 lessons of AI literacy content, all with bilingual infrastructure
- THCS content (Grades 6-9) via HCAI Engine with bilingual support
- lesson-agent — 5-step content pipeline (study guide → slides → teaching guide → quiz → interactive). Currently English-only prompts. YAML extension system allows adding new content types without code.
- creator.learnery — Generates content in 7+ languages. PPTX export. Embed route with language parameter.
What to extend for EMI:
- Add
language: "bilingual" mode to lesson-agent prompts (output Vietnamese content with English integration at configurable ratios)
- Create CLIL YAML prompt flows:
clil-math.yaml, clil-science.yaml — input Vietnamese curriculum standard, output bilingual lesson package
- Use existing HCAI Engine bilingual flag for new EMI subject content (Math, Science)
- Leverage creator.learnery’s multi-language generation for schools that want to create their own CLIL materials
Effort: Medium — new prompt templates, not new systems. The bilingual infrastructure is already there.
4. “Teacher proficiency assessment and development” (KH 18, QD 2371)
What we have:
- PD-Slides — Complete teacher training system:
- 8 modules (3 Foundation + 4 Topic + 1 Implementation), 6-7 hour pathway
- Live quiz engine with timed questions, leaderboard, score-based progression
- Live polling and whiteboard collaboration
- Breakout room activities with configurable groups
- Session timer and activity timers
- Participant registration with session codes and QR access
- Module completion tracking with facilitator approval
- Bilingual certificate generation (EN/VI) with public verification URL
- Google Sheets backend for data sync
- 25+ peer-reviewed research citations backing the pedagogy
- Already has EMI school variant differentiation in activities
- Offline-capable (pure static HTML/CSS/JS)
- Credits-Management — Teacher training certification pipeline:
TrainingModule model with levels (FOUNDATIONAL/METHODOLOGY/COURSE_SPECIFIC)
TrainingCompletion with scoring, section breakdowns, pass/fail, cert status workflow (PENDING → APPROVED → ISSUED)
TeacherQualification model linking teachers to certified modules
Certificate model (unified teacher PD + student course certs) with certCode, verification URL, PDF URL
- 12 API endpoints: submit results, verify, list modules, create/update modules, list/review completions, issue certs, get/revoke qualifications
- Planned: qualification gating on class assignments (teacher must be certified to teach a course)
- AI-Course-fresh admin — PD session management (create sessions, assign modules, enrollment, analytics, certificate issuance/revocation)
What to extend for EMI:
- Create EMI-specific PD modules: “Teaching Math in English”, “CLIL Methodology”, “Classroom English for Subject Teachers”
- The module infrastructure is already there — just new content in the existing PD-Slides format
- Add English proficiency self-assessment module (mapped to VNEFR 6-level framework) — a new quiz module in the existing quiz engine
- Wire the existing
TrainingModule → TeacherQualification pipeline to gate EMI teaching assignments
- Use existing PD certificate system for EMI teacher credentials
Effort: Small — new content modules, not new systems. The entire training + certification pipeline exists.
5. “CLIL (Content and Language Integrated Learning)” (Core EMI methodology)
What we have:
- AI-Course-fresh already implements CLIL concepts:
- Teaching Math and Science concepts through AI literacy content
- Activity types: sorting, matching, MCQ, fill-blank, flashcard, sequence, memory game, drag-classify, detective game, discovery cards, poster design, project showcase
- 25+ interactive activity HTML files
- HCAI Engine slide types: sorting, MCQ, poll, evaluation, timer-activity, comparison-table, vocabulary, flow, process, timeline — all usable for CLIL
- lesson-agent — 7 interactive activity types already defined. YAML extension for new types.
- LMS — Embeds any HTML content (block type “HTML code”), so CLIL activities built anywhere can be delivered through LMS
What to extend for EMI:
- New lesson-agent YAML flows for CLIL subjects (Math, Science, Social Studies)
- Repurpose existing activity types for English vocabulary and subject-specific language practice
- Add subject vocabulary glossary generation step to the pipeline (new YAML tool, ~50 lines)
Effort: Small — new content templates using existing activity infrastructure.
6. “Monitoring and reporting” (Required by all EMI policies)
What we have:
- LMS Analytics — Completion rates, grade distribution, weekly engagement, top performers (teacher-scoped)
- Credits-Management — Full platform KPIs, credit usage by org/user/day, AI generation type breakdown, usage summary rollups, finance module with revenue/expense/P&L/runway
- AI-Course-fresh admin — PD analytics (per-module completion rates, session stats), YCCD compliance tracking (element-by-element coverage heatmap)
- School Admin Portal (Credits-Management) — Self-service portal for school admins: user management, class management, usage stats, content-type breakdown, CSV export, import pipeline
What to extend for EMI:
- Add EMI-specific metrics to LMS analytics: English exposure hours, CLIL lesson completion rates, teacher EMI certification status
- Extend School Admin Portal with EMI readiness scorecard (use existing KPI card pattern)
- Add auto-generated DoE compliance report (new page/export in Credits-Management using existing data)
- Surface existing
report/reportbatch schemas in LMS as actual pages
Effort: Small-Medium — new dashboard views and report templates using existing data infrastructure.
7. “School onboarding and management” (Scaling EMI across schools)
What we have:
- Credits-Management School Onboarding Pipeline — Already built:
- Parse contract PDF (AI-extracted school name, value, dates)
- Parse Excel roster (smart column detection for email/class/name/role)
- Full execution: create org → create users → create classes → assign credits → upload contract to Google Drive
- Bubble sync (bi-directional org/user/class management)
- Temp password issuance (bulk, per-org, with CSV export)
- Multi-org role mapping
- Sales CRM (Credits-Management) — Leads, contacts, activity log, follow-ups, pipeline stages, revenue forecast, link-to-org
- Partner Management — Partner CRUD with linked orgs and purchase orders
- Purchase Orders + Invoicing — Full PO lifecycle (DRAFT → SIGNED), per-class billing, semester billing cycles, invoice generation
What to extend for EMI:
- Add “EMI Pilot” as a lead source / pipeline stage in the CRM
- Create EMI-specific onboarding template (pre-configured org settings, CLIL course assignments, PD module enrollment)
- Bundle the Hanoi Standardized Kit documents into the onboarding flow
Effort: Minimal — configuration, not development.
8. “Mobile access for students” (Expanding reach)
What we have:
- Mobile app — Capacitor 8 (iOS + Android), React/Vite frontend, Express backend, Drizzle ORM, Azure deployment, GitHub Actions CI/CD
What to extend for EMI:
- The mobile app serves as the student-facing EMI practice tool
- Same content delivered through LMS is accessible on mobile
- No additional development needed unless EMI-specific offline content is required
Effort: None for basic EMI delivery.
| EMI Need |
Approach |
What Exists |
What’s New |
Effort |
| AI-powered English tools |
Extend |
ai-gateway, AI chatbots, AI quiz gen |
New chatbot personas, CLIL prompts |
S-M |
| Online learning platform |
Configure |
LMS (full teacher+student) |
EMI course templates, org-level analytics |
S |
| Bilingual content repository |
Extend |
Bilingual infrastructure, HCAI Engine, lesson-agent |
bilingual mode for lesson-agent, CLIL YAML flows |
M |
| Teacher training & certification |
Extend |
PD-Slides (8 modules), training pipeline, certs |
New EMI PD content modules, English self-assessment |
S |
| CLIL activities |
Extend |
25+ activity types, 7 interactive formats |
Subject-specific activity templates |
S |
| Monitoring & reporting |
Extend |
Analytics, admin dashboard, school portal |
EMI metrics, DoE report template |
S-M |
| School onboarding |
Configure |
Full onboarding pipeline, CRM, invoicing |
EMI pipeline stage, onboarding template |
Min |
| Mobile student access |
Use as-is |
Capacitor mobile app |
None |
None |
| EMI Readiness Assessment |
New (small) |
Quiz engine, YCCD compliance tracking |
School self-assessment questionnaire |
S |
| School document kit |
Use as-is |
Hanoi Standardized Kit (4 DOCX templates) |
Adapt for HCMC and other provinces |
Min |
Legend: Min = days, S = 1-2 weeks, M = 2-4 weeks, S-M = varies by scope
The Real Net-New Builds (Only 2)
1. CLIL Content Templates (Medium effort)
New YAML prompt flows for lesson-agent + new bilingual mode. This is the core differentiator — the ability to take any Vietnamese curriculum standard and produce a bilingual CLIL lesson package.
2. EMI Readiness Assessment (Small effort)
A self-assessment tool for schools using the existing PD-Slides quiz engine or LMS assignment system. Maps school capabilities against the 3-level EMI framework from QD 2371.
Everything else is extending what’s already built.
Architecture: How It Fits Together for EMI
┌─────────────────────────────────────────────────────┐
│ SCHOOL ONBOARDING │
│ Credits-Mgmt: CRM → Contract → Org → Users/Classes │
│ + Hanoi Kit (DOCX templates for DoE submission) │
└──────────────┬──────────────────────────────────────┘
│
┌──────────▼──────────┐
│ TEACHER TRAINING │
│ PD-Slides: EMI PD │
│ modules (CLIL, EMI │
│ methodology, English │
│ self-assessment) │
│ → Certificate issued │
│ → TeacherQualification│
└──────────┬──────────┘
│
┌──────────▼──────────┐
│ CONTENT CREATION │
│ lesson-agent: CLIL │
│ YAML flows → bilingual│
│ lesson packages │
│ creator.learnery: │
│ custom CLIL slides │
└──────────┬──────────┘
│
┌──────────▼──────────┐
│ CONTENT DELIVERY │
│ LMS: courses, assign-│
│ ments, quizzes, AI │
│ apps, analytics │
│ AI-Course-fresh: │
│ interactive ebooks, │
│ HCAI Engine (bilingual)│
│ Mobile: student app │
└──────────┬──────────┘
│
┌──────────▼──────────┐
│ AI INFRASTRUCTURE │
│ ai-gateway: GPT-4o, │
│ Claude, Gemini, image│
│ gen, credit billing │
│ Chatbot personas for │
│ English practice │
└──────────┬──────────┘
│
┌──────────▼──────────┐
│ MONITORING & ADMIN │
│ LMS Analytics │
│ School Portal │
│ Credits-Mgmt admin │
│ EMI Readiness tool │
│ DoE compliance report│
└─────────────────────┘