No bio yet
No website
Design and operate scalable PyTorch + Ray distributed training infrastructure for RL workloads at Preference Model to help close the gap between models and real-world use cases.