SmartDine: AI-Powered Restaurant Assistant

SmartDine is an advanced, agentic Retrieval-Augmented Generation (RAG) system designed to simulate a restaurant waiter. Built using LangChain, LangGraph, and OpenAI's GPT models, this project showcases how Generative AI can enhance user experience in daily-life scenarios such as dining and customer support.

---

🚀 Project Objective

To build an intelligent, conversational restaurant assistant that:

• Answers customer queries about the menu.
• Recommends dishes based on user input.
• Provides friendly and context-aware responses.
• Retrieves accurate menu-related content from a custom knowledge base.

---

📊 Tech Stack

Technology	Purpose

Python	Primary language
LangChain	LLM orchestration and tool integration
LangGraph	Workflow and agent logic management
OpenAI GPT	Language understanding and generation
FAISS/ChromaDB	Vector store for retrieval
Jupyter Notebook	Development & experimentation

---

💡 Key Features

1. Agentic Workflow (LangGraph)

• Multi-step dialogue management
• Branching logic and tool usage decisions

2. Retrieval-Augmented Generation (RAG)

• Semantic search of custom documents (e.g., menu, FAQs)
• Injects relevant context into GPT prompt for grounded answers

3. Prompt Engineering

• Custom prompts to emulate a friendly, helpful restaurant waiter
• Ensures consistent tone and persona

4. Memory & Context Handling

• Maintains conversation history
• Personalized and continuous interactions

---

🛍️ Use Case: Daily Restaurant Assistant

SmartDine can be used by:

• Restaurants wanting a virtual assistant on their kiosk/web platform
• Food delivery services adding GenAI chat support
• Digital menu providers aiming to add interactivity
• Smart devices (voice assistants) for dining experiences

---

📅 Sample Interaction

User: What’s special in the vegan section today?

SmartDine: Today’s vegan special is the Grilled Tofu Bowl with quinoa, fresh veggies, and house-made tahini dressing. Would you like a recommendation for something spicy?

---

🔧 How It Works

1. User Input: Text or voice query from customer
2. Intent Detection: GPT interprets user intent
3. Knowledge Retrieval: Relevant docs retrieved using embeddings
4. Response Generation: Context + intent combined and passed to GPT
5. Output: Friendly, coherent waiter-style response

---

🔢 Setup Instructions

1. Clone the repo

git clone https://github.com/yourusername/smartdine-agent.git
cd smartdine-agent

2. Create virtual environment

python -m venv venv
source venv/bin/activate  # For Windows: venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Add your API keys

• Modify .env file with:

OPENAI_API_KEY=your_key_here

5. Run the notebook or script

jupyter notebook Agentic_RAG.ipynb

---

🚀 Future Enhancements

• Voice integration with Whisper or Google STT
• Multilingual support for global audiences
• Integration with live menu APIs
• Real-time user feedback loop

---

👥 Author

Saurabh Kumar Generative AI Engineer LinkedIn | GitHub

---

✅ License

This project is open-source and available under the MIT License.