Introduction
A learning intelligence platform focused on smarter academic workflows, analytics, and personalized digital learning experiences.
Back to Projects
AI Education Platform · Case Study
AI-powered student productivity platform
Role
Full-Stack & AI
Core Stack
Next.js 16 · FastAPI · MongoDB
AI Pipeline
Gemini + Groq LPU
Status
Production Ready
Architecture Overview
SystemComponentArchitecture
Tasks
Schedules
Materials
Assistant
Search
EduSense Core
Quizzes
Analytics
Groups
Auth
Notifications
Full Architecture Breakdown
Engineered for scale, speed & intelligence.
Frontend Architecture
Next.js 16 (App Router) · React 19 · TypeScript
Modular, component-driven client architecture optimized for high interactivity, streaming markdown rendering, and reactive dashboards with zero-layout-shift UI.
Tailwind CSS
Design system & tokens
Recharts
Progress & score charts
Framer Motion
Smooth layout transitions
React Markdown
AI response formatting
Lucide Icons
Vector glyph library
SSR / Streaming
Instant route prefetching
FastAPI REST Engine
Python 3.11 · Modular Routers
Asynchronous endpoint orchestration with Pydantic validation, structured exception boundaries, and automatic interactive documentation.
Modular Router EndpointsSwagger & ReDoc active
POST/authGET/usersPOST/tasksGET/schedulesPOST/documentsPOST/chatGET/groupsPOST/meetingsGET/quiz-scores
AI Engine
Gemini + Groq LPU
Live AI
Dual-engine AI pipeline: ultra-low latency interactive generation via Groq alongside comprehensive concept analysis via Google Gemini.
Smart study schedule & quiz generation
Automated document summary & concept extraction
MongoDB & Beanie ODM
Motor (Async) · 13 Models
Asynchronous NoSQL data models with strict schema typing via Beanie ODM and high-throughput vector queries.
UsersTasksStudy SchedulesStudy MaterialsExtracted ConceptsPDF VectorsStudy GroupsInvitationsMeetingsChat HistoryStudy PlansSessionsQuiz Scores
Auth & Verification
Firebase + Playwright
Authentication Layer
JWT cookie sessions paired with Firebase Authentication and Firebase Admin SDK token validation.
Playwright E2E Testing
End-to-end automated test suites running across Chromium, Firefox, and WebKit browsers.
Scale & System Metrics
Metric 1
10+
Micro-Services Orchestrated
FastAPI modular router architecture
Metric 2
13
Async ODM Data Models
MongoDB with Motor & Beanie ODM
Metric 3
2
Dual AI Engines
Google Gemini + Groq LPU pipeline
Metric 4
100%
Cross-Browser Test Suite
Playwright (Chromium, Firefox, WebKit)
Engineering Insights
Key Engineering Challenges & Solutions
A transparent look at the architectural bottlenecks encountered during development and the engineering decisions used to solve them.
01AI Latency & Precision
The Challenge
"Low-latency AI responses with high structured accuracy."
Architectural Solution
Orchestrated Groq for instant, sub-second token streaming alongside Google Gemini for deep concept extraction, semantic summarization, and smart quiz generation.
02Database & Session Scale
The Challenge
"Scalable session and document vectors."
Architectural Solution
Engineered an asynchronous MongoDB data layer leveraging Motor and the Beanie ODM, ensuring non-blocking vector searches, real-time chat history, and high-throughput student study plans.
03Cross-Browser Reliability
The Challenge
"Flawless multi-user sync and workflow consistency."
Architectural Solution
Implemented end-to-end automated testing pipelines with Playwright across Chromium, Firefox, and WebKit engines to validate critical authentication, real-time study sessions, and responsive UI flows.
Up Next in Portfolio · Commodity Market Intelligence · Haycarb PLC
CarbonXInsight→
Market-analytics web dashboard for coconut shell charcoal prices (US$/MT) across Sri Lanka, India, and Indonesia. Ingests 35+ monthly WPU reports and Excel sheets with automated pdfplumber table extraction and interactive Highstock analytics.
PythonFastAPIReact 19HighchartsMongoDBpdfplumber