Alpha Pizza — AI Generated Smart Deal Finder 🍕⚡

Demonstrating the Business & Customer Value of Gemini 2.5 Flash Multi-Agent Evolutionary Deal Optimization.
Comparing static 20% legacy coupons vs. dynamic, inventory-aware AI-generated heuristics.

---

📸 Application Screenshot

---

🚀 Quickstart & Run Instructions

Prerequisites

• Node.js: v18+ and npm
• Python: 3.11+ and pip
• Google Cloud Access: Vertex AI credentials / gcloud authenticated with access to Gemini 2.5 Flash.

---

1. Clone & Setup Frontend (React + Vite)

# Clone repository
git clone https://github.com/guruvittal/AlphaPizza.git
cd AlphaPizza

# Install dependencies
npm install

# Start Vite frontend dev server (runs on http://localhost:5173)
npm run dev

---

2. Setup & Start Python FastAPI Backend (Gemini 2.5 Flash)

In a second terminal window:

# Navigate to project root
cd AlphaPizza

# Install Python backend dependencies
pip install fastapi uvicorn google-genai pydantic

# Start FastAPI server on port 8000
python3 -m uvicorn server.app:app --host 0.0.0.0 --port 8000

Note: The Vite dev server is configured with a proxy targeting http://localhost:8000/api. When you click "Evolve Deals Live", the app sends requests directly to the Python backend powered live by Gemini 2.5 Flash via Vertex AI.

---

⚡ Architecture Overview

 ┌─────────────────────────────────────────────────────────────────────────────┐
 │ LAYER 1: Background Evolutionary AI Loop (Gemini 2.5 Flash via Vertex AI)   │
 │ • Mutates candidate Python heuristic code based on cart & surplus inventory.│
 │ • Evaluates multi-objective Pareto fitness score F(H).                      │
 └──────────────────────────────────────┬──────────────────────────────────────┘
                                        │ Hot-Swaps Winning Code
                                        ▼
 ┌─────────────────────────────────────────────────────────────────────────────┐
 │ LAYER 2: Real-Time POS & E-Commerce Evaluator (Sub-15ms Checkout SLA)       │
 │ • Instant cart & store inventory evaluation.                                │
 │ • Attaches overstocked store items (Jalapeño Bites, Cinnamon Pull-Aparts,   │
 │   Garlic Knots, Pepsi) at promoted price points.                            │
 └─────────────────────────────────────────────────────────────────────────────┘

---

📊 Core Features

1. 5-Store Regional Network Switcher: Simulates store-level inventory surplus feeds across Midtown, Downtown, Westside, Airport, and Campus locations.
2. Side-by-Side Deal Comparison: Compares legacy static 20% coupons vs. AI Generated Win-Win deals.
3. Multi-Objective Candidate Evaluation Matrix: Ranks all candidate offers by Customer Savings, Store Net Margin ($), and Surplus Stock Attachments.
4. Master Offers & Bundles Catalog: Browse, inspect generated Python code, and switch active strategies across all generations (Gen 0 through Gen 42+).
5. Business ROI Calculator: Simulates network-wide annual profit expansion generated by AI deals across 3,500+ store locations.

---

🛠️ API Reference

Health Check

GET /api/health Returns live backend model status, Vertex AI project ID, and Gemini connectivity.

Live Evolution Trigger

POST /api/evolve Accepts activeCart, storeInventory, currentGeneration, and weights. Calls Gemini 2.5 Flash to dynamically mutate Python deal heuristics and returns the generated candidate strategy object.

---

📄 License

Apache 2.0