encoach_frontend_new_v2/docs/UTAS_MASTER_PLAN.md

EnCoach -- UTAS University Deployment Master Plan

Document Version: 1.0 Date: March 16, 2026 Classification: Confidential -- EnCoach & UTAS Internal Prepared by: EnCoach Engineering Team

---

Table of Contents

1. Executive Summary
2. Project Scope and Objectives
3. System Architecture
4. Feature Modules Breakdown
5. Development Phases
6. University Integration Plan
7. Master Timeline
8. Testing and Rollout Strategy
9. Team and Responsibilities
10. Risk Register
11. Assumptions and Prerequisites

---

1. Executive Summary

EnCoach is deploying a comprehensive AI-powered education platform for the University of Technology and Applied Sciences (UTAS). The platform delivers nine integrated modules spanning self-paced English learning, AI-graded examinations, Mathematics and IT courses with AI-assisted assessment, a full Learning Management System, university SIS integration, institutional whitelabeling, AI virtual tutors, and a support ticketing system.

Platform at a Glance

Dimension	Detail
Modules	9 (English Self Learning, Examinations, Math, IT, LMS, Whitelabeling, Humanisation, SIS Integration, Ticketing)
AI Services	6 (OpenAI GPT-4o, Whisper STT, AWS Polly TTS, ELAI Avatars, GPTZero Detection, FAISS Semantic Search)
Backend	Odoo 19 -- 15 custom modules, 70+ REST API endpoints, PostgreSQL
Frontend	React 18 SPA -- 55+ pages, TypeScript, Tailwind CSS, shadcn/ui
LMS Engine	OpenEduCat (Odoo-native modules)
Team	1 Solution Architect, 1 Senior Developer (AI-assisted), 1 Full-Stack Developer (AI-assisted)

Key Milestones

Milestone	Target Date
Project Kickoff	April 1, 2026
Staging Demo to UTAS	April 24, 2026
English Self Learning + Examinations -- GO LIVE	June 12, 2026
Math + IT + LMS + Ticketing -- GO LIVE	July 24, 2026
Full Platform GO LIVE (SIS, Whitelabeling, Humanisation)	August 21, 2026

Current State of Development

The platform is not starting from zero. Significant engineering has already been completed:

• Backend (Odoo 19): 15 custom modules fully developed, covering exams, AI grading, content generation, media services, training, assignments, classrooms, subscriptions, registration, and ticketing. All AI services (OpenAI, Whisper, Polly, ELAI, GPTZero, FAISS) are integrated and configured. The backend is deployed and running on a staging server.
• Frontend (React): A complete UI prototype with 55+ pages across student, teacher, and admin portals. Role-based routing, layouts, and forms are built. Currently using mock data -- API integration is the primary remaining work.
• Infrastructure: Docker Compose deployment with PostgreSQL, staging environment operational at the current staging server.

---

2. Project Scope and Objectives

2.1 Objectives

1. Deliver a production-ready AI-powered education platform for UTAS covering English, Mathematics, and IT subjects.
2. Integrate with UTAS's existing Student Information System (SIS) for seamless student enrollment, grade sync, and attendance tracking.
3. Provide institutional whitelabeling with UTAS branding, colors, and identity.
4. Deploy AI virtual tutors (Humanisation) to enhance the student learning experience across subjects.
5. Establish a scalable LMS built on OpenEduCat for course management, timetabling, and academic administration.

2.2 Module Scope

#	Module	Description	Acceptance Criteria
1	English Self Learning	AI-powered IELTS preparation: Reading, Listening, Writing, Speaking practice with AI grading, personalized training tips, and progress tracking.	Students can complete full practice sessions in all 4 skills; AI grading returns scores within 60 seconds; training tips are personalized to weaknesses.
2	Examinations	Full exam lifecycle: AI-generated exam content, timed exam sessions, automated grading, results dashboard, and teacher approval workflows.	Teachers can generate, review, and assign exams; students complete timed exams; grading is automated with band scores and feedback.
3	Mathematics	Math courses and AI-assisted exams: course content delivery, formula-based question generation, automated grading for numerical and short-answer responses.	Students access Math courses; AI generates Math questions at configurable difficulty levels; automated grading handles numerical and short-answer formats.
4	Information Technology	IT courses and AI-assisted exams: course content delivery, conceptual and practical question generation, automated grading for MCQ and short-answer formats.	Students access IT courses; AI generates IT questions across topics; automated grading covers MCQ, true/false, and short answers.
5	LMS (OpenEduCat)	Course management, enrollment, timetabling, attendance tracking, batch management, gradebook, and academic reporting via OpenEduCat Odoo modules.	Admin manages courses, batches, and timetables; teachers record attendance and grades; students view schedules and progress.
6	Whitelabeling	UTAS institutional branding: customizable logo, color palette, fonts, login screen, email templates, and report headers.	Platform displays UTAS branding throughout; branding is configurable via admin settings without code changes.
7	Humanisation	AI-powered virtual tutors using ELAI avatar technology: video-based tutoring, interactive speaking practice, and subject-specific AI guidance.	AI avatars present content and interact with students across subjects; avatar selection is available; videos generate within 2 minutes.
8	SIS Integration	Bidirectional integration with UTAS's existing Student Information System: student enrollment sync, grade export, attendance sync, and course catalog mapping.	Student data syncs from SIS to EnCoach on enrollment; grades and attendance export back to SIS; data is consistent within 15-minute sync cycles.
9	Ticketing	Support ticket system for students and staff: ticket creation, categorization, assignment, status tracking, and resolution workflow.	Users submit tickets; staff can categorize, assign, and resolve; status updates are visible to the reporter.

2.3 Out of Scope (This Engagement)

• Mobile native applications (iOS/Android)
• Content creation for Math and IT courses (content to be provided by UTAS or generated via AI)
• UTAS internal network infrastructure changes
• Marketing website / landing page redesign
• Payment gateway integration for UTAS (institutional licensing assumed)

---

3. System Architecture

3.1 Architecture Overview

graph TB
    subgraph frontend ["Frontend Layer"]
        ReactApp["React 18 SPA<br/>TypeScript + Tailwind + shadcn/ui<br/>55+ pages, role-based routing"]
    end

    subgraph backend ["Backend Layer -- Odoo 19"]
        API["REST API Controllers<br/>70+ JSON endpoints"]
        Core["Core Modules<br/>Users, Entities, Roles, Permissions"]
        Exam["Exam Engine<br/>Reading, Listening, Writing, Speaking, Level"]
        LMS["OpenEduCat LMS<br/>Courses, Batches, Timetable, Attendance"]
        AI["AI Modules<br/>Generation, Grading, Media, Training"]
        Sub["Business Modules<br/>Subscriptions, Tickets, Registration"]
    end

    subgraph aiServices ["AI Services"]
        GPT["OpenAI GPT-4o<br/>Content Generation + Grading"]
        Whisper["OpenAI Whisper<br/>Speech-to-Text (local)"]
        Polly["AWS Polly<br/>Text-to-Speech (neural)"]
        ELAI["ELAI<br/>AI Avatar Video"]
        GPTZero["GPTZero<br/>AI Writing Detection"]
        FAISS["FAISS + SentenceTransformers<br/>Semantic Search"]
    end

    subgraph dataLayer ["Data Layer"]
        PG["PostgreSQL 16"]
    end

    subgraph external ["External Systems"]
        SIS["UTAS SIS<br/>Student Information System"]
    end

    ReactApp -->|"JWT Auth + REST API"| API
    API --> Core
    API --> Exam
    API --> LMS
    API --> AI
    API --> Sub
    AI --> GPT
    AI --> Whisper
    AI --> Polly
    AI --> ELAI
    AI --> GPTZero
    AI --> FAISS
    Core --> PG
    Exam --> PG
    LMS --> PG
    Sub --> PG
    API <-->|"REST API Sync"| SIS

3.2 Technology Stack

Layer	Technology	Details
Frontend	React 18, TypeScript, Vite 5	SPA with TanStack Query, React Router v6, Zod validation
UI Framework	shadcn/ui, Tailwind CSS	Radix UI primitives, responsive design, dark mode capable
Backend	Odoo 19, Python 3.11	15 custom modules, ORM-based data access, background job processing
Database	PostgreSQL 16	Odoo-managed schema, full-text search, JSONB for flexible data
LMS	OpenEduCat	Odoo-native education modules: courses, batches, attendance, timetable
AI -- Text Generation	OpenAI GPT-4o, GPT-3.5-turbo	Exam content generation, grading, evaluation, training tips
AI -- Speech-to-Text	OpenAI Whisper (local, base model)	4 instances, round-robin, 30s chunked transcription
AI -- Text-to-Speech	AWS Polly (Neural)	11 English voices, MP3 output, sentence-boundary chunking
AI -- Video	ELAI	7 AI avatars (male/female), ElevenLabs + Azure voices
AI -- Detection	GPTZero	AI-generated writing detection per submission
AI -- Search	FAISS + all-MiniLM-L6-v2	Semantic similarity search over training tips knowledge base
Deployment	Docker Compose	Odoo + PostgreSQL containers, Nginx reverse proxy
Version Control	Gitea	Private repositories for frontend and backend

3.3 Deployment Architecture

graph LR
    subgraph production ["Production Server"]
        Nginx["Nginx<br/>Reverse Proxy<br/>SSL Termination"]
        FrontendContainer["Frontend Container<br/>React SPA<br/>Port 3000"]
        OdooContainer["Odoo 19 Container<br/>Python 3.11<br/>Port 8069"]
        DBContainer["PostgreSQL 16<br/>Port 5432"]
    end

    Browser["Browser"] -->|"HTTPS"| Nginx
    Nginx -->|"/app/*"| FrontendContainer
    Nginx -->|"/api/*"| OdooContainer
    OdooContainer --> DBContainer
    FrontendContainer -->|"REST API"| OdooContainer

3.4 Existing Assets Inventory

Asset	Count	Status
Odoo custom modules	15	Deployed on staging
Backend REST API endpoints	70+	Functional, English/IELTS focused
Frontend pages (React)	55+	UI complete, mock data only
AI service integrations	6	Configured and tested
ELAI avatar profiles	7	Seeded in database
Training tips (FAISS)	80	Imported with embeddings
Odoo data models	20+	Schema deployed

---

4. Feature Modules Breakdown

4.1 English Self Learning

Aspect	Detail
Current State	Backend complete: AI-powered generation for Reading, Listening, Writing, Speaking. Grading via GPT-4o with IELTS rubric. Polly TTS for listening audio. ELAI avatars for speaking. FAISS-based personalized training tips. Frontend UI built (mocked).
Work Required	Replace mock authentication with Odoo JWT. Wire all student-facing pages to backend APIs. Integrate AI grading display (scores, feedback, AI detection). Connect training/tips module.
Dependencies	Stable Odoo API (Phase 1 complete). Frontend auth working.
Effort	6 person-weeks
Risk	Low -- both backend and frontend exist; integration is the main task.

4.2 Examinations

Aspect	Detail
Current State	Backend complete: exam CRUD, 5 module types (reading, listening, writing, speaking, level), AI generation endpoints, approval workflows. Frontend: exam editor, exam list, generation page, approval workflows (all mocked).
Work Required	Wire admin exam management pages to APIs. Connect AI generation UI to real endpoints. Implement exam-taking flow with timer and state persistence. Connect grading pipeline and results display.
Dependencies	English Self Learning module (shared components).
Effort	Included in English Self Learning estimate (shared codebase).
Risk	Low -- core exam engine is proven.

4.3 Mathematics

Aspect	Detail
Current State	No backend module exists. No frontend pages exist. The AI exam engine (GPT-based generation + grading) provides a reusable foundation.
Work Required	Backend: Create encoach_math Odoo module with Math-specific exam models, AI prompt templates for formula-based questions, numerical grading logic, and multi-format question types (MCQ, fill-in, short answer, calculation). Frontend: Create Math course pages, Math exam taking interface with formula rendering (KaTeX/MathJax), Math-specific grading display.
Dependencies	Core exam engine must be stable. AI prompts require testing with GPT-4o for mathematical accuracy.
Effort	5 person-weeks
Risk	Medium -- AI generation of math questions requires careful prompt engineering for accuracy. Formula rendering on frontend adds complexity.

4.4 Information Technology

Aspect	Detail
Current State	No backend module exists. No frontend pages exist. Similar to Math, the exam engine provides a foundation.
Work Required	Backend: Create encoach_it Odoo module with IT-specific exam models, AI prompt templates for conceptual and practical IT questions (networking, databases, programming concepts, security), grading logic for MCQ and short-answer. Frontend: Create IT course pages, IT exam interface, IT-specific grading display.
Dependencies	Core exam engine must be stable.
Effort	4 person-weeks
Risk	Low-Medium -- IT subjects are text-based (no formula complexity). AI generation of IT questions is straightforward with GPT-4o.

4.5 LMS (OpenEduCat)

Aspect	Detail
Current State	OpenEduCat modules will be provided and installed into Odoo 19. Frontend LMS pages exist and are fully built: student dashboard, courses, course detail, assignments, grades, attendance, timetable (student); courses, course builder, assignments, attendance, students, timetable (teacher); dashboard, courses, students, teachers, batches, batch detail, timetable, reports, settings (admin). All pages use mock data.
Work Required	Backend: Install OpenEduCat modules. Create REST API controller layer (encoach_lms_api) to expose OpenEduCat models as JSON endpoints matching the frontend's data expectations. Map OpenEduCat models to frontend data interfaces (Course, Module, Lesson, Batch, Timetable, Attendance, Grade). Frontend: Replace mock data imports with API calls using TanStack Query. Wire forms to create/update APIs.
Dependencies	OpenEduCat modules provided and compatible with Odoo 19.
Effort	3 person-weeks
Risk	Medium -- OpenEduCat model compatibility with Odoo 19 must be verified. API mapping between OpenEduCat's data structures and the frontend's interfaces requires careful alignment.

4.6 Whitelabeling

Aspect	Detail
Current State	No whitelabeling support. Frontend uses CSS variables for theming (Tailwind + shadcn/ui) with darkMode: ["class"] configured but not activated. No branding configuration.
Work Required	Backend: Create encoach_branding model in Odoo for tenant branding settings (logo, primary color, secondary color, font, favicon, login background, email header). Expose branding API endpoint. Frontend: Create a branding provider that fetches configuration on app load and applies CSS variables dynamically. Configure logo, colors, and fonts from API. Add admin UI for branding management.
Dependencies	UTAS provides brand assets (logo, color palette, fonts).
Effort	2 person-weeks
Risk	Low -- CSS variable-based theming is well-established.

4.7 Humanisation (AI Virtual Tutors)

Aspect	Detail
Current State	ELAI integration exists for speaking exam avatar videos. 7 avatars configured (Gia, Vadim, Orhan, Flora, Scarlett, Parker, Ethan) with ElevenLabs and Azure voices. Currently limited to presenting speaking exam questions.
Work Required	Backend: Extend ELAI integration to support tutoring mode -- avatars explain concepts, provide feedback narration, and guide students through exercises. Create endpoints for on-demand tutoring video generation per subject. Frontend: Build tutor interaction UI -- avatar selector, video player with synchronized content, interactive Q&A flow. Integrate across English, Math, and IT modules.
Dependencies	ELAI API capacity and cost model for increased video generation. Subject-specific scripts/prompts.
Effort	3 person-weeks
Risk	Medium -- ELAI video generation time (1-3 minutes per video) affects real-time tutoring UX. Cost scales with usage.

4.8 SIS Integration

Aspect	Detail
Current State	No SIS integration exists. UTAS has an existing Student Information System that we must integrate with via API.
Work Required	Discovery: Obtain UTAS SIS API documentation, authentication method, and data schemas. Backend: Create encoach_sis Odoo module with SIS sync models, scheduled sync jobs (cron), and conflict resolution. Implement: student enrollment import, grade export, attendance sync, course catalog mapping. Frontend: SIS sync status dashboard for admin. Sync logs and error reporting.
Dependencies	UTAS provides SIS API documentation, test environment access, and a technical contact.
Effort	4 person-weeks
Risk	High -- entirely dependent on UTAS SIS API quality, documentation completeness, and test environment availability. This is the single highest-risk module.

4.9 Ticketing

Aspect	Detail
Current State	Backend complete: encoach_ticket Odoo module with full CRUD, categorization, and status workflow. API endpoints exist (/api/tickets). Frontend: Tickets page exists in admin portal (mocked).
Work Required	Wire frontend tickets page to backend API. Add ticket creation flow for students and teachers. Add ticket assignment and resolution workflow for admin.
Dependencies	None -- self-contained module.
Effort	1 person-week
Risk	Low -- both backend and frontend exist.

4.10 Effort Summary

Module	Person-Weeks	Wave
English Self Learning + Examinations	6	Wave 1
Mathematics	5	Wave 2
Information Technology	4	Wave 2
LMS (OpenEduCat)	3	Wave 2
Ticketing	1	Wave 2
Whitelabeling	2	Wave 3
Humanisation	3	Wave 3
SIS Integration	4	Wave 3
Total	28 person-weeks

With a team of 3 producing approximately 3 person-weeks per calendar week, the raw work is approximately 9-10 calendar weeks. Adding UAT cycles, amendments, and buffer, the overall calendar duration is 20 weeks (April 1 -- August 21).

---

5. Development Phases

The project is structured into three overlapping waves. Each wave follows a build-test-deploy cycle and produces a production-ready deliverable.

5.1 Wave 1 -- Foundation (April 1 -- June 12)

Goal: English Self Learning and Examinations modules live in production with real UTAS users.

This wave transforms the existing backend and frontend from separate prototypes into an integrated, production-ready platform.

Sprint Breakdown

Sprint	Dates	Focus	Deliverables
S1	Apr 1-11	Stabilization & Auth	Fix remaining backend bugs. Replace mock auth with Odoo JWT in React frontend. Set up API service layer. Configure production deployment pipeline.
S2	Apr 12-18	Core API Integration	Wire student dashboard, courses, and profile pages. Implement exam list and exam detail APIs. Connect user management and entity management.
S3	Apr 19-25	Exam Engine Integration	Wire exam-taking flow (reading, listening, writing, speaking). Connect AI generation endpoints to exam editor. Integrate timer, state persistence, and submission. Staging demo to UTAS (Apr 24-25).
S4	Apr 26 -- May 9	AI Features & Grading	Wire AI grading pipeline (writing evaluation, speaking evaluation). Connect TTS audio playback. Connect ELAI avatar videos. Integrate training tips and personalized recommendations. Connect AI detection display.
S5	May 10-23	UAT Round 1	Deploy to UTAS staging. Conduct UAT with UTAS staff and selected students. Collect feedback. Bug fixes and amendments.
S6	May 24 -- Jun 6	UAT Round 2 & Hardening	Address UAT feedback. Performance optimization. Security review. Documentation.
S7	Jun 7-12	GO LIVE	Production deployment. Monitoring setup. Handover to UTAS for English + Examinations. Post-launch support begins.

Wave 1 Exit Criteria

• Students can register, log in, and complete full English practice sessions
• AI grading returns accurate band scores with detailed feedback for writing and speaking
• Teachers can generate, review, approve, and assign exams
• Admin can manage users, entities, classrooms, and assignments
• Training tips are personalized based on exam performance
• Platform is stable under expected UTAS load

5.2 Wave 2 -- Expansion (May 17 -- July 24)

Goal: Mathematics, Information Technology, LMS, and Ticketing modules live in production.

Wave 2 begins during Wave 1's UAT phase, leveraging developers freed from active feature work.

Sprint Breakdown

Sprint	Dates	Focus	Deliverables
S8	May 17-30	Math & IT Backend	Create encoach_math and encoach_it Odoo modules. Develop AI prompt templates for Math question generation (numerical, formula-based, word problems). Develop AI prompt templates for IT questions (conceptual, practical). Implement grading logic for non-essay formats.
S9	May 31 -- Jun 13	OpenEduCat & LMS API	Install and configure OpenEduCat modules in Odoo 19. Create encoach_lms_api controller layer. Map OpenEduCat models to frontend data interfaces. Wire LMS admin pages (courses, batches, timetable, attendance).
S10	Jun 14-27	Math & IT Frontend + LMS Wiring	Build Math exam interface with formula rendering. Build IT exam interface. Wire student and teacher LMS pages. Wire ticketing module.
S11	Jun 28 -- Jul 11	UAT Round 1	Deploy Math, IT, LMS, Ticketing to staging. UAT with UTAS staff. Math AI accuracy testing. LMS workflow validation.
S12	Jul 12-18	Amendments	Address UAT feedback. Fix Math/IT grading edge cases. LMS refinements.
S13	Jul 19-24	GO LIVE	Production deployment of Math, IT, LMS, Ticketing. Post-launch support begins.

Wave 2 Exit Criteria

• Math and IT exams generate accurately via AI with appropriate difficulty levels
• LMS (OpenEduCat) fully functional: courses, enrollment, attendance, timetable, gradebook
• Ticketing system operational for students and staff
• All modules integrated with Wave 1 platform (single sign-on, unified navigation)

5.3 Wave 3 -- Integration (July 1 -- August 21)

Goal: Whitelabeling, SIS integration, and Humanisation modules complete. Full platform go-live.

Wave 3 begins during Wave 2's frontend phase, with SIS discovery starting earlier.

Sprint Breakdown

Sprint	Dates	Focus	Deliverables
S14	Jul 1-11	Whitelabeling + SIS Discovery	Implement branding configuration backend and frontend. Apply UTAS brand assets. Begin SIS API discovery and documentation review with UTAS IT team.
S15	Jul 12-25	SIS Integration Development	Build encoach_sis Odoo module. Implement student enrollment import. Implement grade export. Implement attendance sync. Build admin SIS dashboard.
S16	Jul 26 -- Aug 1	Humanisation	Extend ELAI integration for tutoring mode. Build tutor interaction UI. Generate subject-specific avatar tutoring content.
S17	Aug 2-8	Integration Testing	End-to-end testing of all 9 modules together. SIS sync validation with UTAS test data. Performance testing under load. Security audit.
S18	Aug 9-15	Final UAT	Full platform UAT with UTAS. All modules tested together. Final amendments.
S19	Aug 16-21	Full Platform GO LIVE	Production deployment of complete platform. All 9 modules active. Monitoring and support in place.

Wave 3 Exit Criteria

• Platform displays UTAS branding throughout
• SIS syncs student enrollment, grades, and attendance bidirectionally
• AI virtual tutors available across English, Math, and IT
• All 9 modules function as an integrated platform
• Performance meets production requirements under expected load

---

6. University Integration Plan

6.1 Integration Architecture

sequenceDiagram
    participant SIS as UTAS SIS
    participant Odoo as EnCoach Odoo 19
    participant React as EnCoach Frontend
    participant Student as Student

    Note over SIS,Odoo: Enrollment Sync (SIS → EnCoach)
    SIS->>Odoo: Student enrollment data (scheduled sync)
    Odoo->>Odoo: Create/update student records
    Odoo->>Odoo: Assign to courses and batches

    Note over Student,React: Daily Usage
    Student->>React: Login, take courses, complete exams
    React->>Odoo: API calls (exam, grading, attendance)
    Odoo->>Odoo: Record grades, attendance, progress

    Note over Odoo,SIS: Grade & Attendance Export (EnCoach → SIS)
    Odoo->>SIS: Grade export (scheduled sync)
    Odoo->>SIS: Attendance export (scheduled sync)

6.2 Data Exchange Specification

Data Flow	Direction	Frequency	Format
Student enrollment	SIS → EnCoach	On enrollment + daily sync	REST API / CSV import
Course catalog mapping	Bidirectional	On setup + manual sync	REST API
Exam grades	EnCoach → SIS	After grading + daily batch	REST API
Attendance records	EnCoach → SIS	End of day batch	REST API
Student status changes	SIS → EnCoach	Real-time webhook or daily	REST API

6.3 Authentication Strategy

Phase	Method	Detail
Initial (Wave 1-2)	Odoo-native JWT	Students and staff authenticate directly with EnCoach. Credentials provisioned from SIS enrollment data.
Full Integration (Wave 3)	SSO via SAML/OAuth	If UTAS SIS supports SSO, students authenticate once at UTAS portal and access EnCoach seamlessly. Fallback to JWT if SSO is not available.

6.4 UTAS Responsibilities for Integration

Deliverable	Required By	Impact if Delayed
SIS API documentation	June 15, 2026	Blocks SIS module development; delays Wave 3 by equivalent duration
SIS test environment access	July 1, 2026	Blocks integration testing
SIS technical contact (dedicated)	June 15, 2026	Slows discovery and issue resolution
Brand assets (logo, colors, fonts)	June 15, 2026	Blocks whitelabeling completion
UAT testers (3-5 staff, 10-20 students)	May 1, 2026	Delays UAT cycles; risks quality
Math and IT course content (if not AI-generated)	May 15, 2026	Delays Math/IT module population
Production server provisioning	August 1, 2026	Blocks production deployment

---

7. Master Timeline

7.1 Gantt Chart

gantt
    title EnCoach UTAS Deployment -- Master Timeline
    dateFormat YYYY-MM-DD
    axisFormat %d %b

    section Wave1_Foundation
    S1_Stabilize_Auth              :s1, 2026-04-01, 11d
    S2_Core_API_Integration        :s2, 2026-04-12, 7d
    S3_Exam_Engine_Integration     :s3, 2026-04-19, 7d
    UTAS_Staging_Demo              :milestone, m1, 2026-04-24, 0d
    S4_AI_Features_Grading         :s4, 2026-04-26, 14d
    S5_UAT_Round_1                 :s5, 2026-05-10, 14d
    S6_UAT_Round_2_Hardening       :s6, 2026-05-24, 14d
    S7_English_Exams_GO_LIVE       :s7, 2026-06-07, 5d
    English_Exams_LIVE             :milestone, m2, 2026-06-12, 0d

    section Wave2_Expansion
    S8_Math_IT_Backend             :s8, 2026-05-17, 14d
    S9_OpenEduCat_LMS_API          :s9, 2026-05-31, 14d
    S10_Math_IT_LMS_Frontend       :s10, 2026-06-14, 14d
    S11_UAT_Math_IT_LMS            :s11, 2026-06-28, 14d
    S12_Amendments                 :s12, 2026-07-12, 7d
    S13_Math_IT_LMS_GO_LIVE        :s13, 2026-07-19, 5d
    Math_IT_LMS_LIVE               :milestone, m3, 2026-07-24, 0d

    section Wave3_Integration
    S14_Whitelabeling_SIS_Discovery :s14, 2026-07-01, 11d
    S15_SIS_Integration            :s15, 2026-07-12, 14d
    S16_Humanisation               :s16, 2026-07-26, 7d
    S17_Integration_Testing        :s17, 2026-08-02, 7d
    S18_Final_UAT                  :s18, 2026-08-09, 7d
    S19_Full_Platform_GO_LIVE      :s19, 2026-08-16, 5d
    Full_Platform_LIVE             :milestone, m4, 2026-08-21, 0d

    section PostLaunch
    Post_Launch_Support            :pls, 2026-08-21, 30d

7.2 Key Milestones

#	Milestone	Date	Gate Criteria
M0	Project Kickoff	April 1, 2026	Team assembled, scope agreed, environments ready
M1	UTAS Staging Demo	April 24, 2026	English + Exams functional on staging, basic end-to-end flow
M2	Wave 1 GO LIVE	June 12, 2026	English Self Learning + Examinations in production, 2 UAT rounds passed
M3	Wave 2 GO LIVE	July 24, 2026	Math + IT + LMS + Ticketing in production, UAT passed
M4	Full Platform GO LIVE	August 21, 2026	All 9 modules in production, SIS integration validated, full UAT passed
M5	Post-Launch Support Ends	September 21, 2026	30-day support period complete, handover documentation delivered

7.3 Critical Path

The critical path runs through:

1. Wave 1 auth stabilization (blocks all frontend work)
2. Wave 1 API integration (blocks UAT and all subsequent waves)
3. UTAS SIS API documentation delivery (blocks Wave 3 SIS development)
4. OpenEduCat module compatibility (blocks LMS development)

Any delay on items 1-2 cascades through the entire timeline. Items 3-4 are external dependencies that must be tracked proactively.

---

8. Testing and Rollout Strategy

8.1 Testing Stages

graph LR
    Dev["Developer Testing<br/>Unit + Integration<br/>Continuous"] --> Staging["Staging<br/>System Integration<br/>Per Sprint"]
    Staging --> UAT["UAT with UTAS<br/>Acceptance<br/>Per Wave"]
    UAT --> Perf["Performance Testing<br/>Load + Stress<br/>Pre-Go-Live"]
    Perf --> Security["Security Audit<br/>OWASP Top 10<br/>Pre-Go-Live"]
    Security --> GoLive["GO LIVE<br/>Production<br/>Per Wave"]

Stage	Who	When	Scope
Developer Testing	EnCoach dev team	Continuous	Unit tests, API integration tests, manual smoke tests per feature
Staging Deployment	EnCoach dev team	End of each sprint	Full platform on staging server, regression testing, cross-module integration
UAT (User Acceptance)	UTAS staff + students	2 rounds per wave	Guided test scenarios, real-world usage, feedback collection
Performance Testing	EnCoach dev team	Before each go-live	Concurrent user simulation (target: 200 simultaneous users), API response times (<2s for standard, <60s for AI grading)
Security Audit	EnCoach architect	Before production go-live	Authentication review, authorization checks, API input validation, OWASP Top 10

8.2 UAT Process

Each wave includes two UAT rounds:

UAT Round 1 (2 weeks):

1. EnCoach deploys to staging environment
2. UTAS UAT team receives test accounts and test scenarios document
3. UTAS testers execute scenarios and log issues via shared tracking system
4. EnCoach triages issues daily (critical = same day, high = 2 days, medium = next sprint)
5. Weekly status call between EnCoach and UTAS

UAT Round 2 (2 weeks):

1. All critical and high issues from Round 1 resolved
2. Re-test of failed scenarios
3. Exploratory testing by UTAS testers
4. Go-live readiness assessment
5. Sign-off by UTAS project lead

8.3 Go-Live Checklist (Per Wave)

#	Item	Owner
1	All critical and high UAT issues resolved	EnCoach
2	Performance tests pass (response time, concurrent users)	EnCoach
3	Security audit complete, no critical findings	EnCoach
4	Production environment provisioned and configured	EnCoach + UTAS IT
5	DNS / domain configured (if applicable)	UTAS IT
6	SSL certificate installed	EnCoach + UTAS IT
7	Database backup and restore procedure tested	EnCoach
8	Monitoring and alerting configured	EnCoach
9	Rollback procedure documented and tested	EnCoach
10	UTAS project lead sign-off	UTAS

8.4 Rollback Plan

Each go-live deployment includes a documented rollback procedure:

1. Database snapshot taken immediately before production deployment
2. Previous container images tagged and retained (minimum 3 previous versions)
3. Rollback trigger criteria: >5% error rate, >10s average response time, data integrity issue
4. Rollback execution time: <15 minutes (Docker image swap + database restore)
5. Communication: UTAS IT notified within 5 minutes of rollback decision

8.5 Post-Launch Support

Period	Coverage	Scope
Week 1-2 after each go-live	Daily monitoring, same-day critical fixes	Bug fixes, performance tuning, user support
Week 3-4 after each go-live	Standard monitoring, 2-day response for non-critical	Bug fixes, minor enhancements from feedback
After Full Platform GO LIVE	30-day support period (Aug 21 -- Sep 21)	Full platform support, knowledge transfer to UTAS IT

---

9. Team and Responsibilities

9.1 EnCoach Team

Role	Responsibilities
Solution Architect	System architecture decisions. University integration design. SIS API mapping. Security and performance oversight. Code review. UTAS stakeholder communication. Sprint planning.
Senior Developer (AI-assisted)	Backend development (Odoo modules). AI service integration. Math/IT module AI prompts. SIS integration implementation. Database design. API development.
Full-Stack Developer (AI-assisted)	Frontend development (React). API integration. UI/UX implementation. Whitelabeling. LMS frontend wiring. Testing.

9.2 UTAS Responsibilities

Role	Responsibilities
Project Sponsor	Budget approval. Strategic decisions. Escalation resolution.
Project Lead	Scope sign-off. UAT coordination. Go-live approval. Weekly status reviews.
IT Department	SIS API documentation and test environment. Production server provisioning. Network and DNS configuration. SSL certificates.
Subject Matter Experts	Math and IT course content review. Exam quality validation. Training content verification.
UAT Testers (3-5 staff, 10-20 students)	Execute test scenarios. Log issues and feedback. Re-test after fixes.

9.3 Communication Plan

Activity	Frequency	Participants	Medium
Sprint status report	Weekly (Friday)	EnCoach team → UTAS Project Lead	Email + document
Demo / walkthrough	Bi-weekly	EnCoach team + UTAS stakeholders	Video call + screen share
UAT feedback review	Daily during UAT	EnCoach team + UTAS testers	Shared issue tracker
Escalation call	As needed	Architect + UTAS Project Lead	Video call
Go-live readiness review	Before each go-live	All stakeholders	Meeting

9.4 Escalation Path

Level 1: Developer → Solution Architect (same day)
Level 2: Solution Architect → UTAS Project Lead (1 business day)
Level 3: UTAS Project Lead → UTAS Project Sponsor (2 business days)

---

10. Risk Register

#	Risk	Probability	Impact	Mitigation
R1	UTAS SIS API documentation delayed or incomplete	High	High	Begin SIS discovery in Wave 2 (early July). Prepare alternative sync methods (CSV import/export) as fallback. Weekly check-ins with UTAS IT from June onwards.
R2	Math AI question generation inaccuracy	Medium	Medium	Invest in thorough prompt engineering with GPT-4o. Build validation test suite with known-correct Math problems. Include human review step in exam approval workflow.
R3	OpenEduCat incompatibility with Odoo 19	Medium	High	Test OpenEduCat installation early (Sprint S9). Identify compatibility issues before committing to LMS frontend work. Fallback: build minimal LMS models within existing Odoo modules.
R4	Frontend API integration slower than estimated	Medium	Medium	Prioritize critical paths (auth, core pages) first. Use API contract testing to parallelize frontend and backend work. Leverage AI coding assistants to accelerate repetitive wiring.
R5	Team bandwidth constraints	Medium	High	AI-assisted development (both developers use AI tools) effectively increases output by 1.5-2x. Strict scope management -- no feature creep. Clear sprint goals with daily standups.
R6	ELAI video generation latency for Humanisation	Low	Medium	Pre-generate common tutoring videos. Cache frequently requested content. Set user expectations (1-3 min generation time). Consider alternative: pre-recorded avatar videos for standard content.
R7	Production infrastructure not ready on time	Low	High	Request production server provisioning by August 1 (3 weeks before go-live). Test deployment procedure on staging first. Have cloud-based fallback plan.
R8	UTAS UAT testers unavailable or insufficient feedback	Medium	Medium	Provide detailed test scenarios with step-by-step instructions. Offer guided UAT sessions. Establish minimum UAT participation requirements in the scope agreement.

---

11. Assumptions and Prerequisites

11.1 Assumptions

#	Assumption	Impact if Invalid
A1	UTAS provides 3-5 staff and 10-20 students for UAT testing	Delays UAT; reduces quality assurance
A2	OpenEduCat modules are compatible with Odoo 19	Requires building custom LMS models; adds 3-4 weeks
A3	UTAS SIS exposes REST APIs for enrollment, grades, and attendance	Requires alternative integration (CSV, database sync); adds 2-3 weeks
A4	Math and IT course content is either provided by UTAS or generated by AI	If neither, course content creation becomes a separate workstream
A5	The platform will be accessed via web browsers only (no mobile app)	Mobile app would be a separate project
A6	Institutional licensing model (no per-student payment processing)	Payment gateway integration would add 2-3 weeks
A7	Single UTAS deployment (not multi-university / multi-tenant initially)	Multi-tenant support would add architectural complexity
A8	Internet connectivity at UTAS is sufficient for AI cloud services (OpenAI, ELAI, AWS Polly)	On-premise AI deployment would be a separate project

11.2 Prerequisites (with Deadlines)

#	Prerequisite	Owner	Deadline	Status
P1	Scope agreement signed	UTAS + EnCoach	March 31, 2026	Pending
P2	OpenEduCat modules provided	EnCoach / UTAS	May 15, 2026	Pending
P3	UTAS brand assets delivered	UTAS	June 15, 2026	Pending
P4	SIS API documentation provided	UTAS IT	June 15, 2026	Pending
P5	SIS test environment access granted	UTAS IT	July 1, 2026	Pending
P6	UAT test team identified	UTAS	May 1, 2026	Pending
P7	Production server provisioned	UTAS IT / EnCoach	August 1, 2026	Pending
P8	Math/IT course content available	UTAS SMEs	May 15, 2026	Pending

---

Appendix A: Module-to-API Mapping (Existing Backend Endpoints)

The following 70+ API endpoints are already built and deployed in the Odoo 19 backend. These form the foundation for Wave 1 frontend integration.

Category	Endpoints	Count
Authentication	/api/login, /api/logout, /api/reset, /api/reset/sendVerification	4
Users	/api/user, /api/users/update, /api/users/list, /api/users/<id>	4
Entities & Roles	/api/entities, /api/roles, /api/permissions	7
Exams	/api/exam, /api/exam/<module>, /api/exam/<module>/<id>, /api/exam/avatars	7
AI Generation	/api/exam/reading/*, /api/exam/listening/*, /api/exam/writing/*, /api/exam/speaking/*, /api/exam/level/generate	8
Media	/api/exam/listening/media, /api/exam/listening/transcribe, /api/exam/speaking/media, /api/transcribe	6
Evaluations & Grading	/api/evaluate/writing, /api/evaluate/speaking, /api/evaluate/interactiveSpeaking, /api/grading/multiple	6
Classrooms	/api/groups, /api/groups/<id>	3
Assignments	/api/assignments, /api/assignments/<id>, /api/assignments/<id>/start, release, archive	8
Sessions & Stats	/api/sessions, /api/stats, /api/statistical	5
Training	/api/training, /api/training/<id>, /api/training/tips, /api/training/walkthrough	4
Subscriptions	/api/packages, /api/stripe, /api/paypal, /api/paymob + webhooks	7
Registration	/api/register, /api/code/<code>, /api/batch_users, /api/make_user	4
Tickets	/api/tickets, /api/tickets/<id>	3
Storage	/api/storage, /api/storage/delete	2
Approvals	/api/approval-workflows, /api/approval-workflows/<id>	3
Total		~81

---

Appendix B: AI Service Configuration Summary

Service	Model	Key Config	Usage
OpenAI GPT	gpt-4o (primary), gpt-3.5-turbo (secondary)	Temperature: 0.1 (grading), 0.7 (generation), 0.2 (summaries). Max tokens: 4,097. JSON response format.	Exam generation, grading, evaluation, training tips
Whisper	base (local, ~1GB)	4 instances, round-robin, 16kHz mono, 30s chunks with overlap	Speaking transcription, audio-to-text
AWS Polly	Neural engine	11 voices, MP3 output, 3,000 char chunks	Listening exam audio generation
ELAI	7 avatars	ElevenLabs + Azure voices, fade_in animation	Speaking exam video, virtual tutoring
GPTZero	v2/predict/text	Per-sentence analysis, confidence scoring	Writing AI detection
FAISS	IndexFlatL2 + all-MiniLM-L6-v2	Top-5 results, 7 category indices, 80 training tips	Personalized training recommendations

---

Appendix C: Frontend Page Inventory (New React Application)

Portal	Pages	Key Features
Student	Dashboard, Courses, Course Detail, Assignments, Grades, Attendance, Timetable, Profile	8 pages
Teacher	Dashboard, Courses, Course Builder, Course Edit, Assignments, Assignment Detail, Attendance, Students, Timetable, Profile	10 pages
Admin (LMS)	Dashboard, Courses, Students, Teachers, Batches, Batch Detail, Timetable, Reports, Settings, Profile	10 pages
Admin (Platform)	Platform Dashboard, Assignments, Exams List, Exam Structures, Rubrics, Generation, Approval Workflows, Users, Entities, Classrooms, Student Performance, Stats Corporate, Record, Vocabulary, Grammar, Payment Record, Tickets, Settings, Exam	19 pages
Auth	Login, Register, Forgot Password	3 pages
Total		~50+ pages

---

Document prepared by EnCoach Engineering Team. For questions or clarifications, contact the Solution Architect.