Website • Slack • Docs

Scale compute-intensive serverless workloads

Cortex is a Kubernetes-based serverless platform built for AWS.

Deploy realtime, batch, and async workloads

• Realtime - realtime APIs respond to requests in real-time and autoscale based on in-flight request volumes.
• Batch - batch APIs run distributed and fault-tolerant batch processing jobs on-demand.
• Async - async APIs process requests asynchronously and autoscale based on request queue length.

Scale across hundreds of CPU and GPU instances

• No resource limits - allocate as much CPU, GPU, and memory as each workload requires.
• No cold starts - keep a minimum number of API replicas running to ensure that requests are handled in real-time.
• No timeouts - run workloads for as long as you want.

Control your AWS spend

• Spot instance management - Cortex automatically runs workloads on spot instances and falls back to on-demand instances to ensure reliability.
• Multi-instance type clusters - choose the ideal EC2 instance type for your workloads or mix and match several instance types in the same cluster.
• Customizable autoscaling - optimize the autoscaling behavior for each workload to ensure efficient resource utilization.