Technology

Under the Hood

The Wallaroo.AI platform has been purpose-built from the ground up for performance, simplicity, and security

Built for performance, speed, and simplicity

Containerization vs. Wallaroo-ization
Built from ground-up in rust so you are getting C speed Learn more
Kubernetes to go in any cloud, on prem, distributed networks, and the cloud
Installed software so meets your security needs – you own your data at all times
API, SDK or UI for managing models in production
Autoscale: Scaling the resources allocated to each ML pipeline inside Wallaroo.AI based on the specific load for each model or inference Learn more
Observability, logging and auditing

Wallaroo.AI Architecture

Centralized Ops + Distributed Serving

Wallaroo.AI is installed software that goes into any Kubernetes environment, including any cloud, on-prem, and edge.

Leverage our API, SDK, or UI as a central pane of glass to simply deploy, test, observe, and manage models in production.

Your Data Science /
ML Environments

Wallaroo
Operations Center

Your Cloud /
On Prem

Wallaroo
Server

Your Cloud /
On Prem

Wallaroo
Server

Remote Cloud /
On Prem

Wallaroo
Server

Edge
Locations

Wallaroo
Server

Wallaroo.AI Performance

More inferences on less compute

We built our inference serving engine from the ground-up in Rust to deliver more inferences on less compute, saving you cloud costs or enabling deployment in constrained edge environments. Autoscale the resources allocated to each ML pipeline inside Wallaroo.AI based on the specific load for each model or inference.

Learn more

Containerization vs. Wallaroo.AI

The common approach to deployment requires several customer steps for each model in production. Wallaroo.AI provides a purpose-built engine with a highly optimized server that accepts text or binar inputs and supports features like hot-swapping, edge deployments, and more.