Skip to content
JJathuja Sithamparanathan
Back to Projects Archive
STATUS: Completed
INDEX // 2024Featured Engineering Project

TravelHub

Multi-Role Travel Booking & Management Platform

Engineering RoleSoftware Developer / Team Project
Timeline / Year2024
Core DomainReact & TypeScript
Repository StatusCompleted
SCHEMATIC //travelhub.sys_topology.v1

TravelHub Architectural Diagram

TravelHub is a full-stack travel booking and management platform developed as a university group project. My main contributions were the Hotel Owner module and AI chatbot.

NODE_01
React
NODE_02
TypeScript
NODE_03
Spring Boot
NODE_04
PostgreSQL
NODE_05
FastAPI
NODE_06
LangChain
01
THE PROBLEM STATEMENT

Challenge & Ecosystem Context

CORE BOTTLENECK SOLVED

TravelHub is a full-stack travel booking and management platform developed as a university group project. My main contributions were the Hotel Owner module and AI chatbot.

TravelHub is a full-stack travel booking and management platform developed as a university group project. My main contributions were the Hotel Owner module and AI chatbot.

PROJECT CONSTRAINTS
•Zero-downtime fault tolerance
•Strict type safety & validation
•Minimal network serialization overhead
•Cross-client responsive fidelity
02
SYSTEM CAPABILITIES

Key Architectural Features

02.1 // CAPABILITY

End-to-end Hotel Owner Module with admin-approval workflow, room/amenity management, and Supabase Storage image uploads.

SUBSYSTEM MODULEACTIVE
02.2 // CAPABILITY

Comprehensive dashboard providing real-time occupancy insights and rating analytics.

SUBSYSTEM MODULEACTIVE
02.3 // CAPABILITY

AI travel assistant powered by a RAG pipeline (LangChain, Groq LLM, ChromaDB) with intent detection for live backend queries.

SUBSYSTEM MODULEACTIVE
02.4 // CAPABILITY

Event-driven auto-sync service and real-time USD/LKR currency conversion capabilities.

SUBSYSTEM MODULEACTIVE
02.5 // CAPABILITY

Spring Security (JWT) authentication with role-based access for tourists, agencies, hotel owners, and admins.

SUBSYSTEM MODULEACTIVE
03
CRITICAL DIFFERENTIATOR

Challenges & Engineering Solutions

HIGH-CONCURRENCY ARCHITECTURAL TRADE-OFF
VERIFIED PRODUCTION RESOLUTION
The Bottleneck / Obstacle

System Failure Mode Under Peak Load

Integrating diverse system architectures (Java Spring Boot and Python FastAPI) while ensuring real-time consistency and reliable natural language querying against a dynamic relational database.

RISK: RACE CONDITIONS & UI FREEZES
The Engineered Solution

Architectural Strategy & Resolution

Leveraged a Retrieval-Augmented Generation (RAG) pipeline combined with an event-driven auto-sync service, allowing the AI to query real-time data seamlessly across disparate microservices.

OUTCOME: 100% DETERMINISTICVERIFIED
ARCHITECTURAL TOPOLOGY NOTES:

Spring Boot handles core booking and management logic with Spring Security (JWT) and Spring Data JPA; FastAPI serves the AI assistant and ML services. Data is persisted in PostgreSQL with Flyway migrations, vector embeddings in ChromaDB, and media assets in Supabase Storage.

04
QUANTITATIVE RESULTS

Performance & Benchmarks

AI Pipeline
RAG / LangChain
VERIFIED IN BENCHMARK SUITE
Currency Sync
Real-time
VERIFIED IN BENCHMARK SUITE
LLM
Groq LLM
VERIFIED IN BENCHMARK SUITE
Vector DB
ChromaDB
VERIFIED IN BENCHMARK SUITE
05
SPECIFICATION

Tech Stack Composition

ReactTypeScriptSpring BootPostgreSQLFastAPILangChainChromaDBSupabase

Components and modules engineered strictly with type safety, minimal dependency footprints, and production-grade build pipelines.

06
DEPLOYMENT & ARTIFACTS