AI-Powered Web Application using Amazon Bedrock & ECS Fargate
Project Overview
This AI-powered web application leverages Amazon Bedrock and Anthropic’s Claude 3 Haiku model to automatically generate image descriptions for any uploaded image. The entire project is containerized and deployed using AWS ECS Fargate, making the system highly scalable and secure.
Technologies Used
- AWS ECS Fargate: For running containers in a serverless environment, ensuring scalability and high availability.
- Amazon Bedrock: For AI model integration using the Claude 3 Haiku model.
- CI/CD Pipeline: Built using AWS CodeBuild and CodePipeline, with source control managed through GitHub.
- Docker: Containerized the web application, optimized for amd64 architecture to handle deployment challenges.
- IAM Roles: Implemented to secure the deployment process.
Key Highlights
- Infrastructure: Built using AWS ECS Fargate and VPC for secure and scalable cloud-native solutions.
- Containerization: Developed Docker images and optimized them for deployment, overcoming architecture compatibility issues with an M2 MAX chip Mac.
- Automation: The deployment process is fully automated using CI/CD pipelines integrated with GitHub, ensuring smooth updates and version control.
Challenges & Solutions
- Architecture Compatibility: Overcame challenges with deploying on different architectures by creating optimized Docker images for amd64.
- CI/CD Pipeline Setup: Successfully integrated GitHub with AWS CodeBuild and CodePipeline to automate deployments, ensuring all updates are securely and efficiently rolled out.
Lessons Learned
This project provided significant insights into containerization, CI/CD, and deploying AI models on cloud infrastructure. It highlighted the importance of understanding architecture compatibility and how to efficiently manage a continuous deployment process using AWS services.