Touristique
Intelligent Travel. Authentic Experiences.
Created on 2nd November 2025
•
Touristique
Intelligent Travel. Authentic Experiences.
The problem Touristique solves
Touristique addresses information overload, fragmented sources, and lack of personalization in travel planning. It solves poor visibility of local attractions, static recommendations that ignore real-time factors like weather and crowds, language barriers for regional travelers, and budget constraints by delivering AI-powered, context-aware, multilingual itineraries tailored to individual preferences and constraints.
Challenges we ran into
Getting the RAG pipeline working was tricky. Setting up ChromaDB embeddings, indexing the tourism data, and ensuring relevant context reached the LLM required several iterations.
Integrating multiple LLM models was challenging. Balancing local Ollama models (phi3, qwen, llava) with API fallbacks needed careful error handling and prompt engineering to keep responses consistent.
Frontend performance was an issue with high-resolution images causing lag on the profile setup page. We implemented lazy loading, memoization, and CSS optimizations to smooth it out.