SUMM AI serverless AWS architecture
AI startup that simplifies language using Transformer MLP models https://summ-ai.com/ Transformation of an existing MariaDB, Django and ML architecture into a serverless AWS architecture.
AWS, ETL Pipelines, Tensorflow, SQL, Django, Docker, AWS ECS, AWS Security, AWS API Gateway
The goal was to convert the existing Python/MariaDB/ML architecture into a serverless format and transfer it autoscaling and securely to AWS. The existing architecture has not been secure, and did not scale. The individual components were isolated, packaged into Docker images, and deployed into private VPC’s using AWS ECS and API Gateway to ensure maximum security of the “secret” models. The architecture was written in Terraform and is deployed using a Github Actions CI/CD pipeline. The architecture complies with the AWS “well-architected” specification.