Whisker & Wander

Intro/Brochure
Gallery
Exhibit Information
Luna's Chatbot
Memory Graph
Memory Graph 2
Memory Graph 3
Memory Graph 4

🐈 What is Whisker & Wander?

Interactive Virtual Museum - A 3D gallery where users explore exhibits of historical women and their impact across fields like art, technology, and medicine.
User-Driven Exploration - Ask Luna, our AI usher, as much or as little as you want about each exhibit.
Interest-Based Career Discovery - Your interactions map to potential career paths.
Smart Recommendation Engine - Personalized suggestions based on your behavior, powered by our knowledge map.

🏛️ Why use Whisker and Wander?

Most career tools rely on quizzes or static lists. What they lack is inspiration through real stories.

Instead of scrolling through job descriptions, users step into a virtual museum where each exhibit highlights a woman’s impact and how it connects back to them.

Whisker & Wander helps users discover careers through exploration, storytelling, and curiosity.

💡Our Multi-Agentic Model

Agentic RAG System - Combines vector search, a knowledge graph, and user behavior for context-aware responses.
Data Pipeline

[ Markdown Data ] → [ Chunking ] → [ Embeddings ] → [ Chroma DB ]
                           │
                           ▼
              [ Knowledge Graph (NetworkX) ]
                           │
                           ▼
               [ FastAPI + AI Agent Layer ]
                           │
                           ▼
         [ Chat • Search • Recommendations ]

Hybrid Storage - Chroma (vectors) + NetworkX (relationships) + SQLite (user data).
Smart Retrieval - Dynamically switches between vector, graph, and hybrid search.
Personalized Recommendations - Combines graph similarity, user behavior, and importance scoring.
Real-Time Learning - Continuously improves from user feedback.

📜 Framework and Tech Stack

React + Tailwind CSS - Frontend (React Three Fiber for 3D).
FastAPI - Backend + API layer.
Pydantic AI + Gemini API + Chroma - Tool selection + embeddings.
RAG - Context-aware chatbot + recommendations.
NetworkX - Knowledge graph.
SQLite - User sessions + interactions.
Custom Recommendation Algorithm - Graph + behavior + relevance.

🚧 Challenges We Faced

3D Gallery - Building an interactive 3D system with unfamiliar tools.
User Flow - Designing a clean, intuitive experience.
Memory System - Creating a model that learns and updates over time.

🏆 Biggest Accomplishments

Working Knowledge Map - A system that stores, updates, and learns from user interactions in real time.
Hybrid Intelligence System - Successfully combined vector search, graph reasoning, and behavioral learning.
3D Museum Experience - A fully interactive and immersive frontend.
Empowering Decision-Making - Helps women (and all users) explore real stories, build knowledge, and make more confident career decisions.

🚀 Future Plans

Expansive Integrations - Apply our system to any domain with patterns and relationships.
Increased Community - Connect users with shared interests.