Production-Grade AI Recommendation System

Find Your Perfect Smartphone

Powered by advanced AI algorithms combining BM25 retrieval, semantic embeddings, and LLM generation for personalized recommendations.

Live Demo Learn More

Why PerfectPick?

Our AI-powered recommendation system combines multiple advanced technologies to deliver the most accurate and personalized smartphone suggestions.

Hybrid AI Retrieval

Combines classical BM25 with modern BGE embeddings and neural reranking for superior accuracy.

Lightning Fast

Response times under 2 seconds with optimized vector search and caching strategies.

Cloud Native

Deployed on GCP with Kubernetes, Docker, and auto-scaling for enterprise reliability.

Real-time Monitoring

Comprehensive observability with Prometheus and Grafana dashboards.

Session Memory

Personalized recommendations with persistent session history in PostgreSQL.

Highly Scalable

Horizontal Pod Autoscaler handles traffic spikes with 3-10 dynamic replicas.

Multi-LLM Support

OpenAI & Groq integration with fallback mechanisms for reliable generation.

Smart Filtering

Price-based categorization (Budget, Mid-range, Premium, Flagship) with dynamic filtering.

Production Ready

Health checks, metrics endpoints, error handling, and comprehensive logging.

Technology Stack

Built with modern, scalable technologies to ensure high performance and reliability.

Python/Flask
Astra DB
PostgreSQL
Google Cloud
Docker
Kubernetes
Prometheus
Grafana
BM25 Algorithm
BGE Embeddings
Hugging Face
OpenAI/Groq

System Architecture

A modular, cloud-native architecture designed for scalability and performance.

Frontend Layer

HTML/CSS Interface
Responsive Design
User Query Input

Backend Layer

Flask API
Request Handling
Orchestration

AI Processing Layer

BM25 Retrieval
Vector Search
Neural Reranking
LLM Generation

Data Layer

Astra DB (Vector Store)
PostgreSQL (Session Memory)
Flipkart Product Catalog

Monitoring Layer

Prometheus
Grafana
Custom Metrics

Deployment Layer

Docker Containers
Kubernetes (GKE)
Horizontal Pod Autoscaler

API Endpoints

RESTful API designed for easy integration and comprehensive functionality.

GET http://34.14.203.32:8080/ Main Application Interface
POST http://34.14.203.32:8080/recommend Get Personalized Recommendations
GET http://34.14.203.32:8080/health System Health Check
GET http://34.14.203.32:8080/metrics Prometheus Metrics Endpoint

Ready to Experience PerfectPick?

Try our live demo and discover how AI-powered recommendations can transform your smartphone search experience.

Launch Application