PrivateMind isn't just for inference. If your organization has a dedicated GPU allocation, you can submit distributed training jobs to it, work in GPU-backed notebooks, and track every run in MLflow, all inside your own infrastructure, with the same privacy posture as the rest of the platform. Your code, data, and model weights never leave your environment.
What you get
- Distributed training — describe a job once and run it across multiple GPU workers, scheduled onto your org's allocation. Submit it from a notebook, a script, or CI with the Python SDK.
- GPU workspaces — interactive Jupyter notebooks running on your GPUs, with the SDK and common ML libraries preinstalled, wired to your allocation with no setup.
- Experiment tracking — every run streams metrics, parameters, and artifacts to a managed MLflow instance for your org.
How it works
You describe a job (an entrypoint command, a container image, how many workers, and how many GPUs each worker needs) and submit it with the Python SDK. The platform schedules it onto your GPU allocation, runs the distributed job, and streams metrics to MLflow. You get back a handle you can poll, wait on, or cancel.
Inside a PrivateMind GPU workspace, the SDK is preinstalled and preconfigured, so the first cell of a notebook can submit a job with nothing to set up. From a laptop or CI, you authenticate with a PrivateMind API key and name your GPU cluster explicitly.
Start here
- Python SDK — install, authenticate, and submit and manage training jobs from any notebook or script.
