☕ Llama-Latte

Llama-Latte is an AI-powered coffee shop application designed to enhance customer experience through personalized chatbot interactions, real-time recommendations, and seamless order management. Built with React Native and Llama 3 model, it leverages Retrieval-Augmented Generation (RAG) and Market Basket Analysis for accurate, context-aware responses and tailored product suggestions.

---

📸 Screenshots

---

💻 Technologies

Frontend

• React Native
• Expo for rapid app development
• NativeWind for styling

AI and Backend

• Llama 3 Model for AI-driven chatbot
• Retrieval-Augmented Generation (RAG) for personalized responses
• Firebase Firestore for real-time database
• RunPod API for scalable model hosting

Data Science

• Market Basket Analysis using Apriori algorithm for recommendation generation

DevOps

• Docker for containerization
• RunPod for deployment and hosting
• GitHub Actions for CI/CD

---

🚀 Features

🤖 AI-Driven Chatbot

• Built with Llama 3 Model to handle customer queries, recommend products, and manage orders seamlessly.
• Enhanced with RAG for context-aware responses by retrieving real-time data from the coffee shop database.

🛒 Personalized Recommendations

• Utilizes Market Basket Analysis to provide tailored recommendations based on purchase patterns.
• Offers popularity-based and category-specific recommendations for drinks and pastries.

🎯 Agent-Based System

• Specialized agents handle:
  • Order-taking: Validates menu items and calculates totals.
  • Product recommendations: Suggests complementary products.
  • Query filtering: Ensures only relevant questions are processed.

🖥️ Cross-Platform Compatibility

• Developed with React Native for smooth deployment on iOS and Android.
• Deployed on RunPod for scalable API and AI model integration.

---

🛠️ Getting Started

📋 Prerequisites

• Node.js (v14 or higher)
• npm or yarn
• Expo CLI
• Firebase account with Firestore configured
• Docker for local containerization (optional)

🚀 Installation

1. Clone the Repository:

   git clone https://github.com/yourusername/llama-latte.git
   cd llama-latte
   

2. Install Dependencies:

   npm install

3. Set Up Environment Variables: Create a .env file in the root directory with:

   RUNPOD_TOKEN=your_runpod_token

   RUNPOD_CHATBOT_URL=your_runpod_chatbot_url

   MODEL_NAME=meta-llama/Meta-Llama-3-8B-Instruct

   FIREBASE_API_KEY=your_firebase_api_key

4. Start the Application:

   npx expo start --tunnel

📊 Architecture

Frontend

•	React Native app powered by Expo

•	NativeWind for consistent and flexible UI styling

Backend

•	Llama 3 model deployed on RunPod API

•	Firebase Firestore for real-time database management

Recommendation System

•	Apriori algorithm for generating purchase-based recommendations

•	Popularity-based recommendations categorized by product types

🛠️ Key Functionalities

🌟 AI-Driven Order Management

•	Process orders with accurate validation and price calculation.

•	Suggest complementary products based on the customer’s current selection.

🔍 Smart Recommendations

•	Apriori Recommendations: Suggest items frequently purchased together.

•	Popular Recommendations: Suggest trending items in the coffee shop.

•	Category-Specific Recommendations: Suggest items within a specific category based on customer interest.

🔒 Secure and Scalable Deployment

•	Deployed with Docker and RunPod, ensuring scalability and cross-platform compatibility.

•	Sensitive configurations securely managed with environment variables.