Performant Training For All Workstreams Latitude Ai Ray Summit 2026

Summary

GPU utilization is a useful first signal, but past a point it decorrelates from real progress: a busy GPU is not always a fast one. Autonomous trucking is hard. Torc generalizes the task planning stack with reinforcement learning, running two very different RL ... Agents have taken center stage in Agentic production traffic brings long multi-turn sessions, heavy prefix reuse, and bursty, heterogeneous requests that challenge ... Nemotron is one of the leading open model families pushing the frontier of large-scale reinforcement learning. "Github repository: github.com/anyscale/ Miles is an enterprise-facing agentic RL framework for large-scale PyTorch post- Debugging distributed ML jobs on Kubernetes is difficult because of the visibility gap between failures at the infrastructure layer ...