Homepage 2023-08-07T17:49:50+00:00

Serverless compute for machine learning

Run ML jobs on the cheapest GPUs in different clouds

In dynamic environments, you need dynamic infrastructure

Compute resources for machine learning are often virtual machines or Kubernetes clusters in one cloud provider, a setup that historically made sense for many companies and workloads. However, things are changing very quickly in the ML space. It feels like a groundbreaking generative model gets released every few days 🤯 It’s not uncommon to get excited by a model on Hugging Face, only to find it doesn’t run well (or at all) on the GPU you have access to.

Optumi solves this problem by getting you the best GPUs for your use case on-demand, regardless of the cloud provider or datacenter they reside in.

The same experience across clouds

Get the best GPUs for the best prices. Only pay for what you use.

See all jobs in one place

Run on powerful resources in a few lines of Python. You can launch from the Optumi interface or use one of our integrations (e.g. WandB Launch)

Keep track of machines and how much you're spending

Make sure there’s no idle GPUs laying around. Optumi makes it dead simple to see what resources you have running.

View job summaries, logs and notifications on the go

Get texts or emails that update you on job status. You can even configure them on your phone after the job starts running.

See how each resource is being utilized over time

No need to dig through cloud consoles or use DevOps tools. Optumi gives you relevant information in a consumable way.

Feel free to reach out!