AWS Migration
AI DevOps & Infrastructure
Wear It: AI inference migration & image generation
"Working with Tech 42 on our AI Inference Migration was an outstanding experience. Their technical leadership and expertise were instrumental in delivering a scalable AWS infrastructure that powers our content creation and virtual try-on technologies. Our assigned engineer built a robust asynchronous system for generating AI images and enhanced our NoSQL data schema to address the evolving challenges we face as we grow and add new features. What stood out most was his ability to solve complex challenges, reduce errors, and maintain strong project momentum. His deep technical knowledge, effective use of AI automation, responsiveness, and thorough documentation made the entire collaboration seamless and efficient. We are extremely satisfied with the results and excited to build on this solid foundation."
Kitty Yeung
Co-Founder and CEO
Project summary
Wear It needed to migrate their AI workloads from on-premises infrastructure to AWS while scaling to serve millions of outfit visualizations daily. Tech 42 implemented a scalable ECS Fargate architecture with auto-scaling based on queue size. They built an asynchronous image generation system and optimized Wear It's NoSQL data schema to support growth and new features. The infrastructure includes application load balancing for traffic distribution, CloudWatch and Grafana monitoring, and automated CI/CD pipelines for deployments. All infrastructure was defined as code using Terraform. The platform now runs containerized AI models in ECS while maintaining the ability to execute locally during development. Auto-scaling, monitoring dashboards, and CI/CD pipelines were validated and accepted by the customer.
