SmartCart - A Serverless Al-Driven E-Commerce Backend

• 

  SmartCart Architecture Diagram

Inspiration:-

Small online stores often face real challenges when it comes to scaling up their operations. They also deal with issues in personalizing the shopping experience for customers. Analytics can be another tough area to handle properly. Managing the backend infrastructure turns out to be quite costly. It gets complex too, especially for smaller setups. Checkout APIs tend to break down during those high traffic spikes. SmartCart came into being to tackle all these problems. It does so in an effortless way.

What it does :-

SmartCart is a serverless e-commerce backend that automates inventory, checkout, and analytics. It uses Gemini AI Studio to provide smart product recommendations and dynamic pricing. It scales instantly with Google Cloud Run and manages heavy traffic without any downtime. It stores and processes live data using Firestore, Pub/Sub, and BigQuery. It offers secure, test-ready checkout APIs that work with mock payment gateways. It makes online retail intelligent, fast, and easy to maintain.

How we built it :-

Implemented the backend using Node.js (Express) and deployed it on Google Cloud Run for a completely serverless setup. Integrated Firebase Firestore for product, user, and transaction data management. Integrated Gemini AI Studio through the API to provide personalized product recommendations. Used Pub/Sub for event-driven communication and BigQuery for real-time analytics. Implemented mock payment APIs for checkout simulation. Designed CI/CD pipelines using Cloud Build for automated deployment and updates. Designed a minimal frontend (React/Next.js) to test and showcase API integration.

Challenges we ran into

AI Integration: Accurate, timely engineering and fine-tuning responses were necessary to connect the Gemini API with dynamic product data.

** Serverless Optimization:** Keeping latency low while controlling Cloud Run's concurrency limits and cold starts.

** Ensuring real-time updates** between BigQuery, Pub/Sub, and Firestore without causing data loss is known as data synchronization.

** Managing permissions and authentication** in a secure manner when testing in a public setting is known as API security.

** Deployment Pipeline:** Using Cloud Build to set up environment variables and automated builds for smooth CI/CD.

** Resource Limits:** Juggling performance and cost effectiveness within Google Cloud's free-tier restrictions.

Accomplishments that we're proud of

Created an entirely serverless e-commerce backend that requires no manual infrastructure management.

Gemini AI Studio was successfully integrated to provide insightful product recommendations.

High availability and auto-scaling were made possible by Google Cloud Run.

For real-time analytics, a comprehensive data pipeline from Firestore → Pub/Sub → BigQuery was established.

Incorporated secure mock payment APIs into a functional checkout system.

produced a simple, intelligent, and scalable prototype that is ready for production.

What we learned :-

Gained practical experience building serverless architectures with Google Cloud Run and Firebase.

Learned to integrate AI through the Gemini API into real-world backend workflows.

Explored event-driven system design using Pub/Sub and understood its power in creating scalable applications.

Recognized the value of cost optimization and effective concurrency management in cloud-based environments.

Strengthened teamwork, debugging, and version control skills while working under tight development timelines.

Realized how combining AI and cloud technologies can simplify even the most complex e-commerce operations.

Built With

• and-orders.-pub/sub:-event-driven-architecture-for-async-communication.-bigquery:-real-time-analytics-and-data-insights.-artificial-intelligence-gemini-ai-studio-(gemini-api):-personalized-product-recommendations
• cloudbuild
• docker
• firebase
• firebase-auth
• gemini-ai-studio
• google-bigquery
• google-cloud
• google-cloud-run
• node.js
• razorpay
• react.js
• smart-pricing
• users