For data processing backend, Accelex has adopted a microservices-based architecture. Microservices are all containerized and deployed in EKS, AWS’s managed Kubernetes service.
AWS infrastructure is also used to power Accelex’s innovative data analytics pipeline. Data is stored in managed databases (RDS for PostgreSQL) and S3 cloud storage buckets.
Offloading of some rarely used, but resource-intensive, workloads to lambda functions is taking place too.
The data science stack is deployed using lambda functions which allows it to absorb massive spikes in workloads. Sagemaker is also used for ML Ops, including model training, experiment tracking, model registry.