Overview of Project ☁️

Scenario:

Cloudhour, an AI-driven SaaS startup, wants to build a feature that can detect emotions from customer profile pictures and marketing photos.

Their goal is to automatically analyze facial expressions (happy, sad, angry, surprised, etc.) to better understand customer sentiment during campaigns or product feedback sessions.

They need a serverless, real-time AI application that can take an image, analyze it using a pre-trained model, and return the detected emotion instantly, without maintaining any model infrastructure.

Our solution:

We’ll build a fully serverless Image Emotion Detection application that uses the Hugging Face Inference API for AI processing, integrated with AWS Lambda and API Gateway for scalability and automation.

Users will upload an image through a simple web frontend hosted on Amazon S3, which sends it to the API endpoint. Lambda then calls Hugging Face’s facial emotion recognition model and returns the predicted emotion to the browser in real time.

About the Project

In this hands-on lab, you’ll learn to:

• Use the Hugging Face Inference API to perform image-based emotion detection
• Create a Lambda function that connects to Hugging Face securely using API tokens
• Build a REST API using Amazon API Gateway to invoke the Lambda function
• Host a static web frontend on Amazon S3 to upload and test images
• Test the full end-to-end flow from image upload → API → emotion output

By the end of this project, you’ll have a fully functional, AI-powered web app that performs real-time emotion recognition using cloud-native services.

This project demonstrates how to combine Generative AI (Hugging Face) with AWS Serverless Architecture, forming a foundation for future AI/ML applications like facial analysis, customer feedback monitoring, and sentiment-driven automation.

Steps To Be Performed 👩‍💻

We’ll complete the following steps in sequence:

1. Sign up on Hugging Face and obtain your Inference API token
2. Create an AWS Lambda function that connects to Hugging Face
3. Integrate Lambda with API Gateway to expose a public endpoint
4. Build and host the application on Amazon S3
5. Test the complete application end-to-end

Each of these steps will be detailed in the following pages with console screenshots, code snippets, and test examples.

Services Used 🛠

• Hugging Face Inference API → Provides pre-trained models for facial emotion recognition
• AWS Lambda → Executes backend code to process images and call Hugging Face
• Amazon API Gateway → Creates a REST API for frontend-to-backend communication
• Amazon S3 → Hosts the static web interface (HTML, CSS, JS) for image uploads
• Amazon CloudWatch (optional) → Logs Lambda execution details for monitoring and debugging

Estimated Time & Cost ⚙️

• Estimated Time: 2 - 3 hours
• Estimated Cost: ~$0 (within free tier for S3, API Gateway, and Lambda)

Note: Hugging Face’s hosted models can be used free under their community plan with limited request rate.

➡️ Architectural Diagram

This is the architecture you’ll build in this project:

➡️ Final Result

Once completed, you’ll have:

• A deployed serverless AI web app that predicts human emotions from images
• Integration between Hugging Face and AWS services
• Real-time API communication between frontend and backend
• A deeper understanding of AI inference pipelines in the cloud

You’ll not only deploy a working model but also learn how AI APIs and Serverless architecture combine to create intelligent, cost-efficient applications.