(01)
Solution
Training
Multi-week training runs that keep their place when a node fails, with checkpoints you can find again.
0
runs lost to a node failure last quarter
Long runs
What it does
Overhead watches every node in a training job. When one drops, the scheduler pauses the run, moves the shard to a healthy node and resumes from the last checkpoint. You get a log entry, not a page at 3 a.m.
What you set
Checkpoint frequency, which pools a run may use, and the most it may spend. Everything else is handled for you.
(02)
Solutions
What teams use it for.
Long runs
Training
Multi-week training runs that keep their place when a node fails, with checkpoints you can find again.
0
runs lost to a node failure last quarter
Serving
Inference
Serve models across regions with one routing table, and see p95 latency per model, per region, by the minute.
212 ms
median p95 across customer fleets
Inventory
Fleet and regions
Every GPU, node and site in one inventory, including the ones you rent and the ones in your own racks.
6.1%
average idle capacity found in week one
Next step
Try it on your own cluster.
Read-only for 30 days, connected with one of our engineers.