TraceOpt is building the infrastructure layer for efficient AI training.

As training workloads become larger, more distributed, and more expensive, teams need better ways to understand what is happening inside a run. We are building software that makes training more transparent, easier to debug, and more efficient to operate.

We are starting with TraceML, our open-source product for PyTorch training observability, giving teams live visibility into timing, memory, bottlenecks, stragglers, and failures.

Our mission is to make AI training more reliable, efficient, and understandable at scale.

TraceOpt

Org chart

Teams

Offices