10-100x faster data cleaning & model training
Optumi is a cloud service that speeds up machine learning notebooks by optimizing code and computing resources
Problems we solve
Data scientists often wait longer than necessary for ML workloads to complete. Speeding up these workloads requires skills in parallel programming and cloud infrastructure telemetry that they should not be required to master. Optumi takes both problems off their hands so they can spend more time developing great models.
How it works
Frequently asked questions and answers
The Optumi cloud service is accessed through a JupyterLab extension that is easily installed via pip. Please reach out to email@example.com for availability on other popular IDE integrations (VSCode, PyCharm, etc.)
Yes, your notebooks and data are encrypted both in transit (between your laptop and the cloud) and at rest in our SaaS platform. You can easily view and remove files retained in the platform.
To get a taste of the Optumi experience you can submit one notebook for free optimization without being a subscriber. We’ll speedup your code, find the optimal machine for it to run on and give you a runtime / cost estimate.
No credit card required!
To assure the best possible experience for customers, there are currently two restrictions. Submitted notebooks must:
- Run to completion. Our expertise is in workload performance optimization and not in your particular ML domain. We will not be able to debug your notebook for you.
- Complete in under 12 hours. This is to make sure we can return an optimized notebook and runtime/cost estimate within one business day.
For the time being, we are waving the subscription fee and only charging for actual computing resources used. You can pay with a credit card.